29433975_FIT5149 S2 Assignment 1
Huan-yun Wang
8/21/2019
version## _
## platform x86_64-apple-darwin15.6.0
## arch x86_64
## os darwin15.6.0
## system x86_64, darwin15.6.0
## status
## major 3
## minor 6.1
## year 2019
## month 07
## day 05
## svn rev 76782
## language R
## version.string R version 3.6.1 (2019-07-05)
## nickname Action of the Toes1. Introduction
1.1 Background
Superconductors are materials that offer no resistance to electrical current. Prominent examples of superconductors include aluminium, niobium, magnesium diboride, cuprates such as yttrium barium copper oxide and iron pnictides. These materials only become superconducting at temperatures below a certain value, known as the critical temperature [nature.com]. The purpose of this project is to predict the critical tempreatures \(Tc\) of a superconductor based on a set of selected features of chemical properties.
1.2 Data Preparation
1.2.1 Libraries
The libraries that will be used in this assignment
library(caTools)
library(caret)
library(dplyr)
library(glmnet)
library(ggfortify)
library(ggplot2)
library(ggthemes)
library(gridExtra)1.2.2 Loading Dataset
We are using the superconduct dataset from the Superconducting Material Database maintained by Japan’s National Institute for Materials Science(NIMS).
It contains 21,263 material records, each of which have 82 columns: 81 columns corresponding to the features extracted and the last 1 column of the observed Tc values. Among those 81 columns, the first column is the number of elements in the material, the rest 80 columns are features extracted from 8 properties (each property has 10 features).
# Load the data
superconductor <- read.csv("./superconduct/train.csv")
# Display the dimensions
cat("The superconductor dataset has", dim(superconductor)[1], "records, each with", dim(superconductor)[2],
"attributes.")## The superconductor dataset has 21263 records, each with 82 attributes.To get an idea on how our data looks like, we called head() and tail() functions to print out the first and last few rows of the dataset.
# first and last few rows of the dataset
print(head(superconductor))
print(tail(superconductor))To get an idea on how our target values Tc distributed in our dataset, we used summary() and hist(). We saw that the distribution is skewed right with an median of 20. All values are > 0, with a maximum at 185.
summary(superconductor$critical_temp)## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 0.00021 5.36500 20.00000 34.42122 63.00000 185.00000hist(superconductor$critical_temp, breaks = 80, main = "Tc", border="grey", col="dimgrey")
1.2.3 Methodology and Data Split
Now, we are going to split our data into training set/validation set/test set for model selection, fitting, and assessment, using the typical 80:10:10 ratio. Our methodology is to fit the model paremeter for any given complexity on our training set. For every fitted model, we are going to assess the performance on the validation set. We then, based on the performace, select the optimal set of tuning parameters. Finally, for that specific resulting model, we assess a notion of the generalization error using our test set.
# first we generate training set and test set
split = sample.split(superconductor$critical_temp, SplitRatio = 0.8)
training_set = subset(superconductor, split == TRUE)
test_set = subset(superconductor, split == FALSE)
# splits test set into validation set and test set
split = sample.split(test_set$critical_temp, SplitRatio = 0.5)
validation_set = subset(test_set, split == FALSE)
test_set = subset(test_set, split == FALSE)
# reveiew splitting result
split <- c("supercondictor","training_set","validation_set","test_set")
ratio <- c("100%", "80%","10%","10%")
num_records <- c(dim(superconductor)[1],dim(training_set)[1],dim(validation_set)[1],dim(test_set)[1])
num_attributes <- c(dim(superconductor)[2],dim(training_set)[2],dim(validation_set)[2],dim(test_set)[2])
data_dim <- data.frame(split, ratio,num_records, num_attributes)
data_dim## split ratio num_records num_attributes
## 1 supercondictor 100% 21263 82
## 2 training_set 80% 17290 82
## 3 validation_set 10% 1544 82
## 4 test_set 10% 1544 821.2.4 RMSE Function
The defined RMSE function below will be used for calculating RMSE for the following analysis
RMSE <- function(predicted, target) {
se <- 0
for (i in 1:length(predicted)) {
se <- se + (predicted[i]-target[i])^2
}
return (sqrt(se/length(predicted)))
}2. Exploratary Data Analysis
2.1. Collinearity between features
To have a general idea of our data, we first group then and generate subsuets and look at them one by one.
From the 8 plots on the correlations of the features below, we can see there are some simillars pattern across all properties. Particularly, below groups seem to always have strong positive correlations: - mean/wtd_mean/gmean - range/std/wtd_std - entropy/wtd_entropy
The relationship makes sense, as these value are derived from one another, so they all depend on each other at some point.
Subsets 1-4: 10 properties of atomatic_mass/fie/atomatic_radius/Density
# Feature 1: atomic_mass
pairs(superconductor[2:11],main = "Relationship between Properties of Atomic Mass",col="dimgrey")
# Feature 2: fie
pairs(superconductor[12:21],main="Relationship between Properties of Fie",col="dimgrey")
# Feature 3: atomic_radius
pairs(superconductor[22:31],main="Relationship between Properties of Atomic Radius",col="dimgrey")
# Feature 4: Density
pairs(superconductor[32:41], main="Relationship between Properties of Density",col="dimgrey")
Subsets 5-8: 10 properties of ElectronAffinity/FusionHeat/ThermalConductivity/Valence
# Feature 5: ElectronAffinity
pairs(superconductor[42:51],main = "Relationship between Properties of ElectronAffinity",col="dimgrey")
# Feature 6: FusionHeat
pairs(superconductor[52:61],main="Relationship between Properties of FusionHeat",col="dimgrey")
# Feature 7: ThermalConductivity
pairs(superconductor[62:71],main="Relationship between Properties of ThermalConductivity",col="dimgrey")
# Feature 8: Valence
pairs(superconductor[72:81], main="Relationship between Properties of Valence",col="dimgrey")
2.2 Correlations Between Properties
Subset 1: Mean of the 8 properties Mean values are also a good place to start, we will first plot out the correlation between pairs of all 9 variables, including number_of_elements and mean values of all 8 properties, to see if there is anything interesting between each pairs of attributes.
subset_mean <- superconductor[,c(2,12,22,32,42,52,62,72,82)]
dim(subset_mean)## [1] 21263 9str(subset_mean)## 'data.frame': 21263 obs. of 9 variables:
## $ mean_atomic_mass : num 88.9 92.7 88.9 88.9 88.9 ...
## $ mean_fie : num 775 766 775 775 775 ...
## $ mean_atomic_radius : num 160 161 160 160 160 ...
## $ mean_Density : num 4654 5821 4654 4654 4654 ...
## $ mean_ElectronAffinity : num 81.8 90.9 81.8 81.8 81.8 ...
## $ mean_FusionHeat : num 6.91 7.78 6.91 6.91 6.91 ...
## $ mean_ThermalConductivity: num 108 172 108 108 108 ...
## $ mean_Valence : num 2.25 2 2.25 2.25 2.25 2.25 2.25 2.25 2.25 2.25 ...
## $ critical_temp : num 29 26 19 22 23 23 11 33 36 31 ...From the plot we can see there’s stronger relationship between the following pairs: - mean_fie and mean_atomic_radius - mean_atomic_mass and mean_Density - mean_atomic_radius and mean_Density
pairs(subset_mean[1:9], main="Mean of Properties",col = "dimgrey")
This subset of features tend to have similar kind of distribution of the target varialble critical_temp.
par(mfrow = c(3, 3))
hist(subset_mean$mean_atomic_mass, breaks = 20, main = "mean_atomic_mass", border="dimgrey", col="dimgrey")
hist(subset_mean$mean_fie, breaks = 20, main = "mean_fie", border="dimgrey", col="dimgrey")
hist(subset_mean$mean_atomic_radius, breaks = 20, main = "mean_atomic_radius", border="dimgrey", col="dimgrey")
hist(subset_mean$mean_Density, breaks = 20, main = "mean_Density", border="dimgrey", col="dimgrey")
hist(subset_mean$mean_ElectronAffinity, breaks = 20, main = "mean_ElectronAffinity", border="dimgrey", col="dimgrey")
hist(subset_mean$mean_FusionHeat, breaks = 20, main = "mean_FusionHeat", border="dimgrey", col="dimgrey")
hist(subset_mean$mean_ThermalConductivity, breaks = 20, main = "man_ThermalConductivity", border="dimgrey", col="dimgrey")
hist(subset_mean$mean_Valence, breaks = 20, main = "mean_Valence", border="dimgrey", col="dimgrey")
hist(subset_mean$critical_temp, breaks = 20, main = "critical_Temp", border="dimgrey", col="dimgrey")
Subset 2: entropy of the 8 properties
subset_entropy <- superconductor[,c(6,16,26,36,46,56,66,76,82)]
dim(subset_entropy)## [1] 21263 9str(subset_entropy)## 'data.frame': 21263 obs. of 9 variables:
## $ entropy_atomic_mass : num 1.18 1.45 1.18 1.18 1.18 ...
## $ entropy_fie : num 1.31 1.54 1.31 1.31 1.31 ...
## $ entropy_atomic_radius : num 1.26 1.51 1.26 1.26 1.26 ...
## $ entropy_Density : num 1.03 1.31 1.03 1.03 1.03 ...
## $ entropy_ElectronAffinity : num 1.16 1.43 1.16 1.16 1.16 ...
## $ entropy_FusionHeat : num 1.09 1.37 1.09 1.09 1.09 ...
## $ entropy_ThermalConductivity: num 0.308 0.847 0.308 0.308 0.308 ...
## $ entropy_Valence : num 1.37 1.56 1.37 1.37 1.37 ...
## $ critical_temp : num 29 26 19 22 23 23 11 33 36 31 ...pairs(subset_entropy[1:9],main="Entropy of Properties",col="dimgrey")
We noticed this subset of features tend to have jagged and skewed distrubtion, except entropy_ThermalConductivity which is more gaussian distributed
par(mfrow = c(3, 3))
hist(subset_entropy$entropy_atomic_mass, breaks = 20, main = "entropy_atomic_mass", border="dimgrey", col="dimgrey")
hist(subset_entropy$entropy_fie, breaks = 20, main = "entropy_fie", border="dimgrey", col="dimgrey")
hist(subset_entropy$entropy_atomic_radius, breaks = 20, main = "entropy_atomic_radius", border="dimgrey", col="dimgrey")
hist(subset_entropy$entropy_Density, breaks = 20, main = "entropy_Density", border="dimgrey", col="dimgrey")
hist(subset_entropy$entropy_ElectronAffinity, breaks = 20, main = "entropy_ElectronAffinity", border="dimgrey", col="dimgrey")
hist(subset_entropy$entropy_FusionHeat, breaks = 20, main = "entropy_FusionHeat", border="dimgrey", col="dimgrey")
hist(subset_entropy$entropy_ThermalConductivity, breaks = 20, main = "entropy_ThermalConductivity", border="dimgrey", col="dimgrey")
hist(subset_entropy$entropy_Valence, breaks = 20, main = "entropy_Valence", border="dimgrey", col="dimgrey")
hist(subset_entropy$critical_temp, breaks = 20, main = "critical_Temp", border="dimgrey", col="dimgrey")
Subset 3: std of the 8 properties
subset_std <- superconductor[,c(10,20,30,40,50,60,70,80,82)]
dim(subset_std)## [1] 21263 9str(subset_std)## 'data.frame': 21263 obs. of 9 variables:
## $ std_atomic_mass : num 52 47.1 52 52 52 ...
## $ std_fie : num 324 290 324 324 324 ...
## $ std_atomic_radius : num 75.2 67.3 75.2 75.2 75.2 ...
## $ std_Density : num 3306 3767 3306 3306 3306 ...
## $ std_ElectronAffinity : num 51.4 49.4 51.4 51.4 51.4 ...
## $ std_FusionHeat : num 4.6 4.47 4.6 4.6 4.6 ...
## $ std_ThermalConductivity: num 169 199 169 169 169 ...
## $ std_Valence : num 0.433 0.632 0.433 0.433 0.433 ...
## $ critical_temp : num 29 26 19 22 23 23 11 33 36 31 ...pairs(subset_std[1:9],main="Standard Deviation of Properties",col="dimgrey")
Interestingly, this subset of features tend to have multimodal distribution.
par(mfrow = c(3, 3))
hist(subset_std$std_atomic_mass, breaks = 20, main = "std_atomic_mass", border="dimgrey", col="dimgrey")
hist(subset_std$std_fie, breaks = 20, main = "std_fie", border="dimgrey", col="dimgrey")
hist(subset_std$std_atomic_radius, breaks = 20, main = "std_atomic_radius", border="dimgrey", col="dimgrey")
hist(subset_std$std_Density, breaks = 20, main = "std_Density", border="dimgrey", col="dimgrey")
hist(subset_std$std_ElectronAffinity, breaks = 20, main = "std_ElectronAffinity", border="dimgrey", col="dimgrey")
hist(subset_std$std_FusionHeat, breaks = 20, main = "std_FusionHeat", border="dimgrey", col="dimgrey")
hist(subset_std$std_ThermalConductivity, breaks = 20, main = "std_ThermalConductivity", border="dimgrey", col="dimgrey")
hist(subset_std$std_Valence, breaks = 20, main = "std_Valence", border="dimgrey", col="dimgrey")
hist(subset_std$critical_temp, breaks = 20, main = "critical_Temp", border="dimgrey", col="dimgrey")
2.3 Principle Component Analysis
In statistics, PCA is an unsupervised method that linearly projects data from a high dimensional space into a lower dimensional space. By maximising the variance of each of the new, uncorrelated dimensions (principal components), we are able to extract most of the underlying structure and relationships inherent to the original raw data.
Now, because we have observed strong multicollinearity in our data, we are now going to try to use PCA as a tool to better understand and visualise the variance in our dataset in lower dimensions.
# Principal Component
pca_model <- prcomp(training_set[,c(1:81)], center = TRUE,scale. = TRUE)
summary(pca_model)## Importance of components:
## PC1 PC2 PC3 PC4 PC5 PC6 PC7
## Standard deviation 5.6123 2.9097 2.77638 2.53431 2.19071 1.7498 1.70915
## Proportion of Variance 0.3889 0.1045 0.09516 0.07929 0.05925 0.0378 0.03606
## Cumulative Proportion 0.3889 0.4934 0.58855 0.66785 0.72710 0.7649 0.80096
## PC8 PC9 PC10 PC11 PC12 PC13 PC14
## Standard deviation 1.59384 1.38788 1.25962 1.21830 1.09061 0.98034 0.89741
## Proportion of Variance 0.03136 0.02378 0.01959 0.01832 0.01468 0.01187 0.00994
## Cumulative Proportion 0.83232 0.85610 0.87569 0.89402 0.90870 0.92056 0.93051
## PC15 PC16 PC17 PC18 PC19 PC20 PC21
## Standard deviation 0.89414 0.79692 0.75704 0.66623 0.62741 0.55123 0.49471
## Proportion of Variance 0.00987 0.00784 0.00708 0.00548 0.00486 0.00375 0.00302
## Cumulative Proportion 0.94038 0.94822 0.95529 0.96077 0.96563 0.96938 0.97241
## PC22 PC23 PC24 PC25 PC26 PC27 PC28
## Standard deviation 0.4847 0.45592 0.40849 0.40012 0.38704 0.36836 0.33874
## Proportion of Variance 0.0029 0.00257 0.00206 0.00198 0.00185 0.00168 0.00142
## Cumulative Proportion 0.9753 0.97787 0.97993 0.98191 0.98376 0.98543 0.98685
## PC29 PC30 PC31 PC32 PC33 PC34 PC35
## Standard deviation 0.32001 0.30682 0.28763 0.27815 0.27542 0.24097 0.23694
## Proportion of Variance 0.00126 0.00116 0.00102 0.00096 0.00094 0.00072 0.00069
## Cumulative Proportion 0.98811 0.98928 0.99030 0.99125 0.99219 0.99291 0.99360
## PC36 PC37 PC38 PC39 PC40 PC41 PC42
## Standard deviation 0.22401 0.21451 0.19797 0.18718 0.18474 0.16197 0.15789
## Proportion of Variance 0.00062 0.00057 0.00048 0.00043 0.00042 0.00032 0.00031
## Cumulative Proportion 0.99422 0.99479 0.99527 0.99570 0.99612 0.99645 0.99676
## PC43 PC44 PC45 PC46 PC47 PC48 PC49
## Standard deviation 0.14466 0.13867 0.13372 0.13220 0.1267 0.12400 0.12103
## Proportion of Variance 0.00026 0.00024 0.00022 0.00022 0.0002 0.00019 0.00018
## Cumulative Proportion 0.99701 0.99725 0.99747 0.99769 0.9979 0.99808 0.99826
## PC50 PC51 PC52 PC53 PC54 PC55 PC56
## Standard deviation 0.11918 0.11260 0.11078 0.10123 0.09882 0.09732 0.09186
## Proportion of Variance 0.00018 0.00016 0.00015 0.00013 0.00012 0.00012 0.00010
## Cumulative Proportion 0.99843 0.99859 0.99874 0.99887 0.99899 0.99910 0.99921
## PC57 PC58 PC59 PC60 PC61 PC62 PC63
## Standard deviation 0.08503 0.08104 0.08027 0.07608 0.07274 0.06782 0.05995
## Proportion of Variance 0.00009 0.00008 0.00008 0.00007 0.00007 0.00006 0.00004
## Cumulative Proportion 0.99930 0.99938 0.99946 0.99953 0.99960 0.99965 0.99970
## PC64 PC65 PC66 PC67 PC68 PC69 PC70
## Standard deviation 0.05960 0.05639 0.05339 0.05090 0.04760 0.04296 0.04072
## Proportion of Variance 0.00004 0.00004 0.00004 0.00003 0.00003 0.00002 0.00002
## Cumulative Proportion 0.99974 0.99978 0.99982 0.99985 0.99988 0.99990 0.99992
## PC71 PC72 PC73 PC74 PC75 PC76 PC77
## Standard deviation 0.03846 0.03683 0.03458 0.02784 0.02489 0.02089 0.01812
## Proportion of Variance 0.00002 0.00002 0.00001 0.00001 0.00001 0.00001 0.00000
## Cumulative Proportion 0.99994 0.99995 0.99997 0.99998 0.99999 0.99999 0.99999
## PC78 PC79 PC80 PC81
## Standard deviation 0.01362 0.01089 0.008585 0.00699
## Proportion of Variance 0.00000 0.00000 0.000000 0.00000
## Cumulative Proportion 1.00000 1.00000 1.000000 1.00000The result tells us that we are actually able to capture up to almost 99% of variance in the entire dataset using only 30 principal components.
Now we will quickly run a Principal Component Regression using some of the key principal components we just calculated and see how it goes.
pcr_model <- train(critical_temp ~ .,
data = training_set,
method = 'pcr',
tuneGrid = expand.grid(ncomp = seq(2,40,2)),
trControl = ,
preProc = c('center','scale','BoxCox'))
pcr_model$results## ncomp RMSE Rsquared MAE RMSESD RsquaredSD MAESD
## 1 2 25.91598 0.4340713 21.12392 0.1870371 0.004371131 0.1352689
## 2 4 23.59378 0.5309428 18.75033 0.1613195 0.004617488 0.1195303
## 3 6 22.25655 0.5825855 17.70229 0.1204943 0.005058058 0.1317304
## 4 8 22.23655 0.5833334 17.59996 0.1227173 0.004968249 0.1066751
## 5 10 22.23127 0.5835297 17.57764 0.1234933 0.004984291 0.1083197
## 6 12 21.90477 0.5956759 17.26656 0.1266269 0.005641423 0.1005386
## 7 14 21.78913 0.5999365 17.14338 0.1259035 0.005517950 0.1016760
## 8 16 21.57609 0.6077168 17.03000 0.1231018 0.005751325 0.1007561
## 9 18 21.44371 0.6125268 16.89100 0.1244964 0.005549185 0.1027890
## 10 20 21.28112 0.6183859 16.83165 0.1214638 0.005731132 0.1062709
## 11 22 20.89875 0.6319685 16.71036 0.1268960 0.005834470 0.1061440
## 12 24 20.58353 0.6429930 16.34391 0.1530739 0.006460945 0.1100342
## 13 26 19.79650 0.6697523 15.58452 0.1639544 0.007013933 0.1201760
## 14 28 19.77439 0.6704892 15.57147 0.1519972 0.006749447 0.1188111
## 15 30 19.51542 0.6790511 15.23949 0.1715930 0.007098828 0.1622978
## 16 32 19.36942 0.6838576 15.02386 0.1792539 0.006888911 0.1141317
## 17 34 19.36122 0.6841234 15.01951 0.1863832 0.007303873 0.1179752
## 18 36 19.30942 0.6858202 14.95558 0.1742180 0.006772224 0.1316568
## 19 38 18.92883 0.6980752 14.50738 0.1655878 0.006702987 0.1187027
## 20 40 18.94355 0.6976208 14.50319 0.1827815 0.007312864 0.1160223pcr_model$bestTune## ncomp
## 19 38Not too well with just an rsqured of 0.7 wiht 40 principal componets. The underfitted model might be the result of a small number of principal components \(d\), or the potential non-linear relationship between the predictors and response variable. Which we will be discussing more in the model selection section.
But what are the significant features identified by this algorithm?
pcr_features = varImp(pcr_model)
pcr_top40 = data.frame(feature = pcr_features$importance%>% rownames(),
overall = pcr_features$importance$Overall)
pcr_top40 = pcr_top40[order(pcr_top40$overall,decreasing = TRUE),][1:40,]
# Generates a slice for top 40 important features
pcrFeatures = pcr_top40$feature %>% as.character()
# generates a subset from superconductor
pcrFeatures <- superconductor %>% select(append(pcrFeatures,"critical_temp"))We plot the histograms of all these features. We noticed that there are 4 features having similar distributions with critical_temp, including:
- wtd_gmean_Density
- wtd_mean_FusionHeat
- wtd_gmean_FusionHeat
- wtd_range_FusionHeat
# log transformation on features with skewed dist
N = ncol(pcrFeatures)
colorcode <- rep("dimgrey",N)
colorcode[N] <- "deeppink"
par(mfrow=c(3, 3))
for (i in 1:(N)) {
hist(pcrFeatures[,i], breaks = 20, main = paste(i,names(pcrFeatures)[i],sep="."), border=colorcode[i],
col=colorcode[i],cex.main=0.9)
}
We further explore some interesting relationships between the significant features through data visualisation.
ggplot(superconductor,
aes(x =wtd_gmean_Density , y =critical_temp, color = wtd_std_ThermalConductivity)) +
geom_point(aes(size =number_of_elements), alpha = 0.4) +
ggtitle("Data Exploration - Figure 1") +
theme(plot.title = element_text(hjust = 0.5)) +
scale_color_distiller(palette = "Paired") +theme_light()
ggplot(superconductor,
aes(x = number_of_elements , y = wtd_gmean_Valence, color = wtd_mean_Valence)) +
geom_point(aes(size = wtd_std_ThermalConductivity), alpha = 0.4) +
ggtitle("Data Exploration - Figure 2") +
theme(plot.title = element_text(hjust = 0.5)) +
scale_color_distiller(palette = "Paired") +theme_light()
ggplot(superconductor,
aes(x = number_of_elements , y =wtd_gmean_Density , color = entropy_atomic_mass)) +
geom_point(aes(size = wtd_std_ThermalConductivity), alpha = 0.4) +
ggtitle("Data Exploration - Figure 3") +
theme(plot.title = element_text(hjust = 0.5)) +
scale_color_distiller(palette = "Paired") +theme_light()
ggplot(superconductor,
aes(x = number_of_elements , y =entropy_atomic_mass, color =critical_temp )) +
geom_point(aes(size = mean_atomic_mass), alpha = 0.4) +
ggtitle("Data Exploration - Figure 4") +
theme(plot.title = element_text(hjust = 0.5)) +
scale_color_distiller(palette = "Paired") +theme_light()
ggplot(superconductor,
aes(x = wtd_entropy_Valence , y =wtd_entropy_atomic_radius , color = wtd_entropy_FusionHeat)) +
geom_point(aes(size = range_atomic_radius), alpha = 0.4) +
ggtitle("Data Exploration - Figure 5") +
theme(plot.title = element_text(hjust = 0.5)) +
scale_color_distiller(palette = "Paired") +theme_light()
ggplot(superconductor,
aes(x = wtd_gmean_Density, y =wtd_gmean_fie , color = wtd_std_fie)) +
geom_point(aes(size = range_fie), alpha = 0.4) +
ggtitle("Data Exploration - Figure 6") +
theme(plot.title = element_text(hjust = 0.5)) +
scale_color_distiller(palette = "Paired") +theme_light()
library(ggthemes)
ggplot(superconductor,
aes(x = wtd_gmean_Density, y =log(critical_temp), color =wtd_std_atomic_radius)) +
geom_point(aes(size = std_atomic_radius), alpha = 0.4) +
ggtitle("Data Exploration - Figure 7") +
theme(plot.title = element_text(hjust = 0.5)) +
scale_color_distiller(palette = "Paired") +theme_light()
ggplot(superconductor,
aes(x = range_fie, y =std_atomic_radius, color =wtd_std_atomic_radius)) +
geom_point(aes(size = wtd_mean_FusionHeat), alpha = 0.4) +
ggtitle("Data Exploration - Figure 8") +
theme(plot.title = element_text(hjust = 0.5)) +
scale_color_distiller(palette = "Paired") +theme_light()
3. Model Developmemnt
3.1 Linear regression
3.1.1 Linear regression with all variables
Our first model is going to be the simplest one, fitting a linear regression model to all of the 81 indicators. We fit the model by calling lm(), and then we call summary() to summarize the results.
# Fitting Simple Linear Regression to the Training set
fit1.1 = lm(formula = critical_temp ~ .,
data = training_set)There are a few points we wanted to highlight here: - The R-squred 0.74 indicates that the model explains 74% of the variation in Tc - The F-statistic 608 has a p-value < 2.2e-16, so reject the null hypothesis (the model explains nothing) and accept the althernative hypothesis that the model is useful
# prints important stats
num_features1.1 <- dim(summary(fit1.1)$coefficients)[1]-1
cat("Number of features in model 1.1 = ",num_features1.1)## Number of features in model 1.1 = 81cat("\nRsquared = ",summary(fit1.1)$adj.r.squared)##
## Rsquared = 0.7394145cat("\nF-statistic =", summary(fit1.1)$fstatistic[1])##
## F-statistic = 606.65133.1.1 Linear Regression with Stepwise Feature Selection
Initially, we had all 81 in our model. But we doubted if they are all that important and necessary. To really find out the optmized subset of features, the all subset algorithms is an option. However, in this case, given the size of our dataset and the number of attibutes in hand, due to the fact that the computational complexity of such brute force algorithm is exponential.
Our other option is to do a sub-optimal approach such as stepwise algorithm for feature selection. How stepwise search (or Greedy search) works is you start with some set of possible features (or zero feature), and then you greedily walk through features, and select the best one to take or drop, and then you keep iterating. Here, we ran the step() function, a stpewise algorithm for feature selection by AIC, to find it out.
Even though this procedure is significantly more computational efficient
O(\(D^{2}\)) >> O(\(2^{D}\)) for large D
given the search data size and search time, we only did backwards and both direcation searches this time.
# Run step to remove unnecessary variables
sback_fit1.1 = step(fit1.1,direction = "backward")
sboth_fit1.1 = step(fit1.1,direction = "both")
# extract AIC
aic_fit1.1 <- extractAIC(fit1.1)
aic_sback_fit1.1 <- extractAIC(sback_fit1.1)
aic_sboth_fit1.1 <- extractAIC(sboth_fit1.1)According to the step() results, the backward/both selection gave us the same results, removing 11 features from our model and achieved lower AIC.
stepResults.fit1 = data.frame(
"num_predictors" =
c("beginning.fit"=aic_fit1.1[1]-1,
"step.backward"=aic_sback_fit1.1[1]-1,"step.both"=aic_sboth_fit1.1[1]-1),
"aic" =
c("beginning.fit"=aic_fit1.1[2],
"step.backward"=aic_sback_fit1.1[2],"step.both"=aic_sboth_fit1.1[2])
)
stepResults.fit1## num_predictors aic
## beginning.fit 81 99149.91
## step.backward 73 99140.21
## step.both 73 99140.21Now let’s prepare to remove these features and update the model
# feature removed by step()
removed <- sback_fit1.1$anova$Step
# string argument for updating the linear model
formula = paste(".~.",paste(removed,collapse = ""),sep = "")
formula## [1] ".~.- wtd_range_Density- wtd_std_ThermalConductivity- wtd_mean_Density- wtd_range_Valence- wtd_range_atomic_mass- wtd_entropy_atomic_mass- wtd_entropy_ThermalConductivity- gmean_atomic_radius"We updated our model 1.1 and removed the 7 features by using the string input we defined above. The updated model gave us a similar rsqured while using less predictors.
# update fit1
fit1.2 <- update(fit1.1,formula)
# prints out stats
num_features1.2 <- dim(summary(fit1.2)$coefficients)[1]-1
cat("Number of features in model 1.2 = ",num_features1.2)## Number of features in model 1.2 = 73cat("\nRsquared = ",summary(fit1.2)$adj.r.squared)##
## Rsquared = 0.73944083.1.2 Linear Regression with Significant Variables
In addition to the step() function above, we wanted to explore some other options in feature selection. Here, we used varImp() function to calculate importance of all 81 predictors initially in our first model.
# train the model
fit1.3 <- train(critical_temp ~., data = training_set, method = 'lm',preProcess="scale",trControl = trainControl(method = "cv"))
# List of features with their importance scores
importance1.3 <- varImp(fit1.3, scale=FALSE)
print(importance1.3)## lm variable importance
##
## only 20 most important variables shown (out of 81)
##
## Overall
## std_ElectronAffinity 18.657
## range_ElectronAffinity 18.390
## wtd_mean_ThermalConductivity 16.263
## wtd_mean_atomic_radius 12.905
## wtd_entropy_FusionHeat 12.432
## wtd_std_ElectronAffinity 12.336
## wtd_range_ThermalConductivity 12.238
## wtd_gmean_atomic_radius 11.654
## wtd_gmean_ElectronAffinity 11.593
## range_atomic_mass 11.526
## wtd_std_Valence 11.127
## wtd_entropy_Valence 10.655
## wtd_gmean_ThermalConductivity 10.402
## range_fie 9.652
## wtd_mean_FusionHeat 9.580
## wtd_mean_ElectronAffinity 9.263
## mean_atomic_mass 9.129
## wtd_entropy_ElectronAffinity 9.117
## wtd_gmean_FusionHeat 8.593
## mean_FusionHeat 8.593Let’s organize the data nad rank the features by importance score and extract the top 60 for modeling
# extracts reletive elements from the original output list
rank = importance1.3$importance$Overall
# generates a new list called features including columns of feature names and their rankings
features <- training_set %>% select(-critical_temp) %>% names()
important1.3 = data.frame(features, rank)
important1.3 = important1.3[order(important1.3$rank,decreasing=TRUE),]
# Generates a slice for top 60 important features
top60 = important1.3[1:60,1] %>% as.character()
top60## [1] "std_ElectronAffinity" "range_ElectronAffinity"
## [3] "wtd_mean_ThermalConductivity" "wtd_mean_atomic_radius"
## [5] "wtd_entropy_FusionHeat" "wtd_std_ElectronAffinity"
## [7] "wtd_range_ThermalConductivity" "wtd_gmean_atomic_radius"
## [9] "wtd_gmean_ElectronAffinity" "range_atomic_mass"
## [11] "wtd_std_Valence" "wtd_entropy_Valence"
## [13] "wtd_gmean_ThermalConductivity" "range_fie"
## [15] "wtd_mean_FusionHeat" "wtd_mean_ElectronAffinity"
## [17] "mean_atomic_mass" "wtd_entropy_ElectronAffinity"
## [19] "wtd_gmean_FusionHeat" "mean_FusionHeat"
## [21] "gmean_FusionHeat" "mean_Density"
## [23] "wtd_range_FusionHeat" "std_atomic_mass"
## [25] "range_atomic_radius" "std_Density"
## [27] "wtd_mean_atomic_mass" "entropy_atomic_mass"
## [29] "wtd_entropy_fie" "std_fie"
## [31] "wtd_entropy_Density" "wtd_entropy_atomic_radius"
## [33] "entropy_FusionHeat" "wtd_range_ElectronAffinity"
## [35] "std_ThermalConductivity" "range_ThermalConductivity"
## [37] "range_Density" "wtd_gmean_atomic_mass"
## [39] "entropy_fie" "gmean_atomic_mass"
## [41] "wtd_range_fie" "entropy_Valence"
## [43] "wtd_range_atomic_radius" "entropy_ThermalConductivity"
## [45] "wtd_std_FusionHeat" "range_Valence"
## [47] "entropy_Density" "entropy_atomic_radius"
## [49] "range_FusionHeat" "number_of_elements"
## [51] "wtd_std_atomic_radius" "gmean_ElectronAffinity"
## [53] "wtd_gmean_Density" "wtd_gmean_Valence"
## [55] "wtd_std_Density" "mean_atomic_radius"
## [57] "std_atomic_radius" "gmean_Valence"
## [59] "wtd_mean_Valence" "std_Valence"Re-fits Top60 Features
This time, with using only 60 selected features, we achieve almost the same r-square 0.74 we obtained using up to 74 features previously. In this case, we would say thay the varImp() did better job than step() in feature selection, because it provides more effiecient solution for our model.
# generates a subset of traninig data for top 60 features
features_top60 <- append(top60,"critical_temp")
training1.3 <- training_set %>% select(features_top60)## Note: Using an external vector in selections is ambiguous.
## ℹ Use `all_of(features_top60)` instead of `features_top60` to silence this message.
## ℹ See <https://tidyselect.r-lib.org/reference/faq-external-vector.html>.
## This message is displayed once per session.# refits model using top60 features
fit1.3 = lm(formula = critical_temp~ .,
data = training1.3)
num_features1.3 <- dim(summary(fit1.3)$coefficients)[1]-1
cat("Number of features in model 1.3 = ",num_features1.3)## Number of features in model 1.3 = 60cat("\nRsquared = ",summary(fit1.3)$adj.r.squared)##
## Rsquared = 0.7371022Assessment on Validation Set
Now that we have three fitted models at the moment, we are going to assess each of them on the validation set and compare their performance.
# Model 1.1 Predicting Tc for training/validation set
pred_train1.1 = predict(fit1.1, newdata = training_set)
rmse_train1.1 <- RMSE(pred_train1.1, training_set$critical_temp)
pred_validation1.1 = predict(fit1.1, newdata = validation_set)
rmse_v1.1 <- RMSE(pred_validation1.1, validation_set$critical_temp)
rsq_v1.1 <- cor(pred_validation1.1, validation_set$critical_temp)^2
# Model 1.2 Predicting Tc for training/validation set
pred_train1.2 = predict(fit1.2, newdata = training_set)
rmse_train1.2 <- RMSE(pred_train1.2, training_set$critical_temp)
pred_validation1.2 = predict(fit1.2, newdata = validation_set)
rmse_v1.2 <- RMSE(pred_validation1.2, validation_set$critical_temp)
rsq_v1.2 <- cor(pred_validation1.2, validation_set$critical_temp)^2
# Model 1.3 Predicting Tc for training/validation set
pred_train1.3 = predict(fit1.3, newdata = training_set)
rmse_train1.3 <- RMSE(pred_train1.3, training_set$critical_temp)
pred_validation1.3 = predict(fit1.3, newdata = validation_set)
rmse_v1.3 <- RMSE(pred_validation1.3, validation_set$critical_temp)
rsq_v1.3 <- cor(pred_validation1.3, validation_set$critical_temp)^2RMSE Analysis
Comparing the three models, we believe that model 1.3 is a better one, since it was able to acheieve the similar results using less predictors.
lin_reg_model1.1 <- c("num_predictors"=num_features1.1,
"adj.rsquared_train"=summary(fit1.1)$adj.r.squared,
"adj.rsquared_validation"=rsq_v1.1,
"rmse_validation"=rmse_v1.1)
lin_reg_model1.2 <- c("num_predictors"=num_features1.2,
"adj.rsquared_train"=summary(fit1.2)$adj.r.squared,
"adj.rsquared_validation"=rsq_v1.2,
"rmse_validation"=rmse_v1.2)
lin_reg_model1.3 <- c("num_predictors"=num_features1.3,
"adj.rsquared_train"=summary(fit1.3)$adj.r.squared,
"adj.rsquared_validation"= rsq_v1.3,
"rmse_validation"= rmse_v1.3)
# creates a df by combing above vectors
models.1 <- data.frame(lin_reg_model1.1,lin_reg_model1.2,lin_reg_model1.3)
models.1## lin_reg_model1.1 lin_reg_model1.2 lin_reg_model1.3
## num_predictors 81.0000000 73.0000000 60.0000000
## adj.rsquared_train 0.7394145 0.7394408 0.7371022
## adj.rsquared_validation 0.6963862 0.6964549 0.6951216
## rmse_validation.3 18.4305527 18.4284397 18.4647374Diagnostic Plots
We will firt call autoplot() from ggfortify package to calculate and produce diagnostic plots and see if there are any serious problems inherented in our model.
- Residuals vs Fitted : shows the residuals are not evenly distributed around zero, this suggests that the assumption that the relationship is linear is not reasonable, also the variances of the error terms are unequal.
- Normal Q-Q indicates the plot tells us there’s evidence of non-linearity
- Scale-Location plot tells us the residuals are not spread equally along the ranges of predictors
- Residuals vs Leverage helps us to find influential cases, but there seems no such case here
autoplot(fit1.3)
Generalization of Error
Since model 1.3 has been selected as our first official model based on the performance on the validation, we are now going to assess its performance on the test set. We then assume the test errors as an approximation of our generalization error.
# Predicting Tc for training/test
pred_test1.3 = predict(fit1.3, newdata = test_set)
rmse_test1.3 <- RMSE(pred_test1.3,test_set$critical_temp)
# Calculates RMSE of training pred
cat("\nLINEAR REGRESSION MODEL 1.3: RMSE for the test predictions =", rmse_test1.3)##
## LINEAR REGRESSION MODEL 1.3: RMSE for the test predictions = 18.46474Visualization of Gernerailzed Error
As we can see, we got this very similar RMSE that we have seen on the training/validation set. To get a better idea on the fitness, we visualized the perfomance using ggplot().
From the plot, we could tell that, althoght not that strong, there’s still a linear relationship between the true values and the predicted ones. That should be a fair representation on the fitness of this model.
# Visualizing the fit
Linear_regression_test <- ggplot() +
geom_point(aes(x = test_set$critical_temp, y = pred_test1.3),
colour = 'gold2',alpha=0.5,size=3) +
ggtitle('Linear Regression') +
ylab('Prediction') +
xlab('True Value (Tc)') +
theme_minimal() +
geom_abline(colour = "grey80", size = 1)
Linear_regression_test 
3.1.3 Linear Regression with Transfored Dataset
To further improve our model, we are going to try and generate two-way interaction terms for fitting our second model. Particularly, we are going to generate interaction terms using features with low collinearity, since we know that collinearity, which is the correlation between predictor variables supply redundant information to the model, and may consequently effect model performance.
Thus, the first step is to pin down these features using the findCorrelation() function. By setting the cutoff threshold as 0.8, we got 29 selected features
Collinearity Analysis
# Identifies highly correlated terms
correlationMatrix <- cor(training_set[,1:81])
# findCorrelation() searches through a correlation matrix
# and returns a vector of integers corresponding to columns to remove to reduce pair-wise correlations.
highlyCorrelated <- findCorrelation(correlationMatrix, cutoff=0.8)
cat("Number of highly correlated features (to be removed) =",length(highlyCorrelated))## Number of highly correlated features (to be removed) = 52# creates a subset containing only features with correlation < 0.8
lowCor = training_set[,-highlyCorrelated]
cat("\nNumber of features with correlation lower than 0.8 (to be selected) =",length(lowCor)-1)##
## Number of features with correlation lower than 0.8 (to be selected) = 29Skewness Analysis
As we looked at the histograms of all 29 features selected based on the collinearity, we found that most of the features have weird distributions.
# looks at distribution of each feature
par(mfrow=c(3, 3))
N = ncol(lowCor) -1
for (i in 1:N) {
hist(lowCor[,i], breaks = 20, main = paste(i,names(lowCor)[i],sep = ". "), border="grey", col="darkgrey")
}
Account for the Heteroscedasticity
Since there are too many features to look at at a time, we automatically generated histograms for those features (those do not have any zero values) with log transformation, and check the distributions again of each one of them.
We spotted the three of them (highlighted in yellow) have become more gaussian in their distributions after log transformation. These following five features below with log transformations will be included in our model:
- gmean_fie
- mean_Density
- gmean_ElectronAffinity
- wtd_gmean_ElectronAffinity
- mean_FusionHeat
# log transformation on features with skewed dist
par(mfrow=c(3, 3))
N = ncol(lowCor) -1
colorcode <- rep("gray87",29)
colorcode[c(3,9,13,14,18)] <- "goldenrod"
for (i in 1:N) {
if (min(lowCor[,i])>0){
hist(log(lowCor[,i]), breaks = 20, main = paste(i,paste("log(",names(lowCor)[i],")",sep = ""),sep = ". "), border=colorcode[i], col=colorcode[i])
}
}
We removed the three features with skewed distributions and added the log transformed terms back into our string argument for fitting the model
# generate a subset of features with low correlation coeficients, while excluding 3 highly skewed terms
features2.1 <- lowCor %>% select(-c(3,9,13,14,18))%>%select(-critical_temp) %>% names()
# adding back the three originally skewed terms after applying log transformation, and generates an string argument for updating the model fit
formula <- paste(paste(features2.1,collapse = "+"),"log(gmean_fie) + log(mean_Density) + log(gmean_ElectronAffinity)+log(wtd_gmean_ElectronAffinity)+log(mean_FusionHeat))^2",sep = "+")
formula <- paste0(".~. +(",formula)
formula## [1] ".~. +(mean_atomic_mass+std_atomic_mass+wtd_gmean_fie+wtd_range_fie+mean_atomic_radius+wtd_range_atomic_radius+std_atomic_radius+wtd_entropy_Density+wtd_range_Density+wtd_std_Density+wtd_entropy_ElectronAffinity+wtd_range_ElectronAffinity+std_ElectronAffinity+wtd_range_FusionHeat+std_FusionHeat+mean_ThermalConductivity+wtd_gmean_ThermalConductivity+entropy_ThermalConductivity+wtd_entropy_ThermalConductivity+wtd_range_ThermalConductivity+std_ThermalConductivity+gmean_Valence+range_Valence+wtd_range_Valence+log(gmean_fie) + log(mean_Density) + log(gmean_ElectronAffinity)+log(wtd_gmean_ElectronAffinity)+log(mean_FusionHeat))^2"Fit the interaction terms (feature crosses) to the model
The rsquare value turned out to be 0.82, much higher than the previous 0.74. However, we have also seen a big jump in the number of features. Maybe some of the features are not that important and thus can be excluded. To reduce the number of features while trying to keep up with the good performance, we are going to do feature selection later.
# refits the model by updating model 1
fit2.1 <- update(fit1.3,formula,data=training_set)
# prints stats to console
num_features2.1 <- dim(summary(fit2.1)$coefficients)[1]-1
cat("MODEL 2.1: number of predictors = ",num_features2.1)## MODEL 2.1: number of predictors = 477cat("\nTraining adj.rsquared = ",summary(fit2.1)$adj.r.squared)##
## Training adj.rsquared = 0.8152777Diagnostic plots
- Residuals vs Fitted plot tells us the residuals have non-linear patterns
- Normal Q-Q indicates the residuals are not normally distributed.
- Scale-Location plot suggests that the residuals are not spread equally along the ranges of predictors, as the residuals do not appear randomly spread
- Residuals vs Leverage gives us a typical look when there is no influential case
autoplot(fit2.1)
Performance of model 2.1 on the validation set was better than our previous models, as we got a lower RMSE here
# Predicting Tc for validation set
pred_v2.1 <- predict(fit2.1, newdata = validation_set)
# Calculates RMSE
rmse_v2.1 <- RMSE(pred_v2.1, validation_set$critical_temp)
cat("\nMODEL 2.1: Validation RMSE =", rmse_v2.1)##
## MODEL 2.1: Validation RMSE = 16.3187# Rsquared
rsq_v2.1 <- cor(pred_v2.1, validation_set$critical_temp)^2
cat("\nValidation adj.rsquared = ",rsq_v2.1)##
## Validation adj.rsquared = 0.7618833.1.4 Linear Regression with Regularisation
As more features we use, the more complex the model becomes, and more likely it becomes overfit. And high complexity models could have low bias, but high variance. In this case, we want to trade off between bias and variance to get to that sweet spot of having good predictive performance.
One way to automatically balance between bias and variance is called regularization. To balance between the two measures, we introduced a new term \(lambda\) and modified the cost function as below:
Total Cost = (Measure of Fit) + \(\lambda\)*(Magnitude of Coefficients)
In essence, the tuning parameter \(\lambda\) controls model complexity, and controls such bias/variance trade-off.
In this section, we are going to use two regulariztion techniques: Ridge Regression and Lasso Regression to fit to our second model.
Data preparation
When we ran the lm() function on the interaction terms, we input this string argument critical_temp ~(.)^2. But it does not work for the glmnet function, as takes matices as input arguments. So we need to manually create the set of features.
To generate the interaction terms, the first step is to put together all the selected features (with low collinearity) as a matrix using model.matrix(). After the interaction terms are generated , we convert the matrix back to dataframe, so we can easily combine the top 60 important features used in model 1 and the interaction terms as the training set for regularization.
# creates a copy of 60 important features used in model 1
training2.2 <- training1.3
validation2.2 <- validation_set %>% select(names(training2.2))
test2.2 <- test_set %>% select(names(training2.2))
# generates subset of features with low correlation while replaces the originally skewed terms with log transformed terms
# generates string argument for later fitting the model.matrix()
features_lowCor <- lowCor %>% select(-c(3,9,13))%>%select(-critical_temp) %>% names()
formula <- paste("critical_temp~(",paste(paste(features_lowCor,collapse = "+"),"log(gmean_fie) + log(mean_Density) + log(gmean_ElectronAffinity))^2",sep = "+"),sep="")
formula## [1] "critical_temp~(mean_atomic_mass+std_atomic_mass+wtd_gmean_fie+wtd_range_fie+mean_atomic_radius+wtd_range_atomic_radius+std_atomic_radius+wtd_entropy_Density+wtd_range_Density+wtd_std_Density+wtd_gmean_ElectronAffinity+wtd_entropy_ElectronAffinity+wtd_range_ElectronAffinity+std_ElectronAffinity+mean_FusionHeat+wtd_range_FusionHeat+std_FusionHeat+mean_ThermalConductivity+wtd_gmean_ThermalConductivity+entropy_ThermalConductivity+wtd_entropy_ThermalConductivity+wtd_range_ThermalConductivity+std_ThermalConductivity+gmean_Valence+range_Valence+wtd_range_Valence+log(gmean_fie) + log(mean_Density) + log(gmean_ElectronAffinity))^2"Prepares training set for regularization…
# creates a matrix for interaction terms using the string argument we created above
train_interact <- model.matrix(critical_temp~(mean_atomic_mass+std_atomic_mass+wtd_gmean_fie+wtd_range_fie+mean_atomic_radius+wtd_range_atomic_radius+std_atomic_radius+wtd_entropy_Density+wtd_range_Density+wtd_std_Density+wtd_gmean_ElectronAffinity+wtd_entropy_ElectronAffinity+wtd_range_ElectronAffinity+std_ElectronAffinity+mean_FusionHeat+wtd_range_FusionHeat+std_FusionHeat+mean_ThermalConductivity+wtd_gmean_ThermalConductivity+entropy_ThermalConductivity+wtd_entropy_ThermalConductivity+wtd_range_ThermalConductivity+std_ThermalConductivity+gmean_Valence+range_Valence+wtd_range_Valence+log(gmean_fie) + log(mean_Density) + log(gmean_ElectronAffinity))^2,training_set)[,-1]
# removes duplicates by indexing and filtering
train_interact <- train_interact %>% as.data.frame()
train_interact <- train_interact[which(!names(train_interact) %in% names(training2.2))]
training2.2 <- cbind(train_interact,training2.2)
dim(training2.2)## [1] 17290 476Same procedure of preparing validation set for regulariztion…
# creates a set of features including interaction terms (test set)
v_interact <- model.matrix(critical_temp~(mean_atomic_mass+std_atomic_mass+wtd_gmean_fie+wtd_range_fie+mean_atomic_radius+wtd_range_atomic_radius+std_atomic_radius+wtd_entropy_Density+wtd_range_Density+wtd_std_Density+wtd_gmean_ElectronAffinity+wtd_entropy_ElectronAffinity+wtd_range_ElectronAffinity+std_ElectronAffinity+mean_FusionHeat+wtd_range_FusionHeat+std_FusionHeat+mean_ThermalConductivity+wtd_gmean_ThermalConductivity+entropy_ThermalConductivity+wtd_entropy_ThermalConductivity+wtd_range_ThermalConductivity+std_ThermalConductivity+gmean_Valence+range_Valence+wtd_range_Valence+log(gmean_fie) + log(mean_Density) + log(gmean_ElectronAffinity))^2,validation_set)[,-1]
v_interact <- v_interact %>% as.data.frame()
v_interact <- v_interact[which(!names(v_interact) %in% names(validation2.2))]
validation2.2 <- cbind(v_interact,validation2.2)
dim(validation2.2)## [1] 1544 476Still the same proceduer preparing test set for regulariztion…
# creates a set of features including interaction terms (test set)
test_interact <- model.matrix(critical_temp~(mean_atomic_mass+std_atomic_mass+wtd_gmean_fie+wtd_range_fie+mean_atomic_radius+wtd_range_atomic_radius+std_atomic_radius+wtd_entropy_Density+wtd_range_Density+wtd_std_Density+wtd_gmean_ElectronAffinity+wtd_entropy_ElectronAffinity+wtd_range_ElectronAffinity+std_ElectronAffinity+mean_FusionHeat+wtd_range_FusionHeat+std_FusionHeat+mean_ThermalConductivity+wtd_gmean_ThermalConductivity+entropy_ThermalConductivity+wtd_entropy_ThermalConductivity+wtd_range_ThermalConductivity+std_ThermalConductivity+gmean_Valence+range_Valence+wtd_range_Valence+log(gmean_fie) + log(mean_Density) + log(gmean_ElectronAffinity))^2,test_set)[,-1]
test_interact <- test_interact %>% as.data.frame()
test_interact <- test_interact[which(!names(test_interact) %in% names(test2.2))]
test2.2 <- cbind(test_interact,test2.2)
dim(test2.2)## [1] 1544 476Before we fit data to the Ridge regression, let’s do not forget to convert our dataframes into matices.
# transforms df to matrix as input args of the glmnet()
xmat_train <- training2.2 %>% select(-critical_temp) %>% as.matrix()
ymat_train <- training2.2 %>% select(critical_temp) %>% as.matrix()
xmat_v <- validation2.2 %>% select(-critical_temp) %>% as.matrix()
ymat_v <- validation2.2 %>% select(critical_temp) %>% as.matrix()
xmat_test <- test2.2 %>% select(-critical_temp) %>% as.matrix()
ymat_test <- test2.2 %>% select(critical_temp) %>% as.matrix()
dim(xmat_train)## [1] 17290 475dim(xmat_v)## [1] 1544 475dim(xmat_test)## [1] 1544 475Fit Training Data to L2 Ridge Regression
Now that the data is ready, we ready to fit the data using 10-fold cross validation to optimise labda for L2 Shrinkage Penalty.
# fit a ridge model with cross-validation using the cv.glmnet() function
ridge2.2 <- cv.glmnet(xmat_train, ymat_train, type.measure="mse",
alpha=0, family="gaussian")
plot(ridge2.2)
Ridge Assessment on Validation Set
Now we will use the predict() function to apply ridge model to the training set and the validation set using the optimimal lambda value.
The optimal \(\lambda\) can be extracted by lambda.1se, which is \(\lambda\)*, the lambda value resulted in the simplest model (the model with the fewest non-zero parameters). We know that the lambda value was within 1 standard error of the lambda that rsulted in the smallest total cost.
predicts on the validation set
# using the optimised lambda to predict
cat("Optimized lambda =",ridge2.2$lambda.1se)## Optimized lambda = 2.965807# Coefficients
coef_ridge2.2 <- coef(ridge2.2)[-1, 1]
coef_ridge2.2 <- coef_ridge2.2[order(abs(coef_ridge2.2), decreasing = TRUE)]
num_coef_ridge2.2 <- length(coef_ridge2.2[coef_ridge2.2 >0])
cat("\nnumber of predictors = ",num_coef_ridge2.2)##
## number of predictors = 245# predicts on training set
pred_train_ridge2.2 <- predict(ridge2.2, s=ridge2.2$lambda.1se,newx=xmat_train)
# MSE
rmse_train_ridge2.2 <- RMSE(pred_train_ridge2.2, training_set$critical_temp)
cat("\nTraining RMSE of Ridge regression on model 2 =",rmse_train_ridge2.2)##
## Training RMSE of Ridge regression on model 2 = 16.56204# Rsquared
rsq_train_ridge2.2 <- cor(pred_train_ridge2.2, training_set$critical_temp)^2
cat("\nTraining adj.rsquared of Ridge regression on model 2 = ",rsq_train_ridge2.2)##
## Training adj.rsquared of Ridge regression on model 2 = 0.7687567predicts on the validation set
# predict on validation set
pred_v_ridge2.2 <- predict(ridge2.2, s=ridge2.2$lambda.1se,newx=xmat_v)
# MSE
rmse_v_ridge2.2 <- RMSE(pred_v_ridge2.2, validation_set$critical_temp)
cat("\nValidation RMSE of Ridge regression on model 2 =",rmse_v_ridge2.2)##
## Validation RMSE of Ridge regression on model 2 = 17.5677# Rsquared
rsq_v_ridge2.2 <- cor(pred_v_ridge2.2, validation_set$critical_temp)^2
cat("\nValidation adj.rsquared of Ridge regression on model 2 = ",rsq_v_ridge2.2)##
## Validation adj.rsquared of Ridge regression on model 2 = 0.7232634Fit Training Data to Lasso Regularisation
Now we move on to fit the data cv.glmnet() to optimise lambda for L1 Shrinkage Penalty.
From the plot blow, we see that as lambda increases, the number of features shrinks, and the mean-squared error increases. When \(\lambda\) approaches zero, we get minimized MSE.
#### alpha = 1, Lasso Regression
################################
# fit a lasso model with cross-validation using the cv.glmnet() function
lasso2.2 <- cv.glmnet(xmat_train, ymat_train, type.measure="mse",
alpha=1, family="gaussian")
plot(lasso2.2)
Similarly, we will fit the model to the training set and the validation set using the optimized lambda value.
# min lambda
cat("Optimized lambda =",lasso2.2$lambda.1se)## Optimized lambda = 0.00624274# coefficients
coef_lasso2.2 <- coef(lasso2.2)[-1, 1]
coef_lasso2.2 <- coef_lasso2.2[order(abs(coef_lasso2.2), decreasing = TRUE)]
num_coef_lasso2.2 <- length(coef_lasso2.2[coef_lasso2.2 >0])
cat("\nnumber of predictors = ",num_coef_lasso2.2)##
## number of predictors = 187# Predict Tc on Training set
pred_train_lasso2.2 <- predict(lasso2.2, s=lasso2.2$lambda.1se, newx=xmat_train)
# MSE
rmse_train_lasso2.2 <- RMSE(pred_train_lasso2.2, training_set$critical_temp)
cat("\nTraining MSE of Ridge regression on model 2 =",rmse_train_lasso2.2)##
## Training MSE of Ridge regression on model 2 = 15.09433# Rsquared
rsq_train_lasso2.2 <- cor(pred_train_lasso2.2, training_set$critical_temp)^2
cat("\nTraining adj.rsquared of Lasso regression on model 2 = ",rsq_train_lasso2.2)##
## Training adj.rsquared of Lasso regression on model 2 = 0.8072066Lasso Assessment on Validation Set
As we predicted on the validation set using the optimized lambda value, we found that Lasso Regularization gave us better results.
# predict on validation set
pred_v_lasso2.2 <- predict(lasso2.2, s=lasso2.2$lambda.1se, newx=xmat_v)
# MSE
rmse_v_lasso2.2 <- RMSE(pred_v_lasso2.2,validation_set$critical_temp)
cat("\nValidation MSE of Lasso regression on model 2 = ",rmse_v_lasso2.2)##
## Validation MSE of Lasso regression on model 2 = 16.58278# R-squared
rsq_v_lasso2.2 <- cor(pred_v_lasso2.2, validation_set$critical_temp)^2
cat("\nValidation adj.r-squared of Lasso regression on model 2 = ",rsq_v_lasso2.2)##
## Validation adj.r-squared of Lasso regression on model 2 = 0.7536771RMSE Analysis
Comparing the three models, we believe that Lasso did a better job in regulariztion given the better r-squared, lower MSE, and fewer number of coefficients.
The result makes sense because Lasso regression is known for excluding useless variables from equations and making the final equation simpler and easier to interpret (and we assumed that most of the generated terms in our second model are not that important). So it is a better than Ridge regression at reducing the variance in models that contain lots of useless predictors. In contrast, Ridge regression tend to do a little better when most variables are useful.
# creates vectors of stats
lin_reg_trans <- c("num_predictors"=num_features2.1,
"adj.rsquared_train"=summary(fit2.1)$adj.r.squared,
"adj.rsquared_validation"=rsq_v2.1,
"rmse_validation"=rmse_v2.1)
lin_reg_ridge <- c("num_predictors"=num_coef_ridge2.2,
"adj.rsquared_train"=rsq_train_ridge2.2,
#"mse_train"=mse_train_ridge2.2,
"adj.rsquared_validation"=rsq_v_ridge2.2,
"rmse_validation"=rmse_v_ridge2.2)
lin_reg_lasso <- c("num_predictors"=num_coef_lasso2.2,
"adj.rsquared_train"=rsq_train_lasso2.2,
#"mse_train"=mse_train_lasso2.2,
"adj.rsquared_validation"= rsq_v_lasso2.2,
"rmse_validation"= rmse_v_lasso2.2)
# creates a df by combing above vectors
models.2 <- data.frame(lin_reg_trans,lin_reg_ridge,lin_reg_lasso)
models.2## lin_reg_trans lin_reg_ridge lin_reg_lasso
## num_predictors 477.0000000 245.0000000 187.0000000
## adj.rsquared_train 0.8152777 0.7687567 0.8072066
## adj.rsquared_validation 0.7618830 0.7232634 0.7536771
## rmse_validation.3 16.3187004 17.5677006 16.5827787Skewness Analysis
Lasso regression achieved 0.77 in rsquared with fewer attributes.
To get a better idea, we took a closer look at the distribution of these important features in the Lasso Model. However, due to the number of features, we only look at those with coefficient > 0.01.
We found that most of the features have weird and wild distributions. Again, let’s try to do log transformation automatically on all of them and see if they look more gaussian
# checks skewness of important features in Lasso model
features <- names(coef_lasso2.2[coef_lasso2.2 >0.01])
features_lasso <- training2.2 %>% select(features)## Note: Using an external vector in selections is ambiguous.
## ℹ Use `all_of(features)` instead of `features` to silence this message.
## ℹ See <https://tidyselect.r-lib.org/reference/faq-external-vector.html>.
## This message is displayed once per session.# plot hist
par(mfrow=c(3, 3))
N = ncol(features_lasso)
for (i in 1:N) {
hist(features_lasso[,i], breaks = 20, main = paste(i,names(features_lasso)[i],sep = ".\n"), border="grey", col="darkgrey")
}
Some of the features became more gaussian in their distribution after log transformation.
- log(gmean_ElectronAffinity)
- log(wtd_gmean_ElectronAffinity:gmean_Valence)
# log transformation on features with skewed dist
N = ncol(features_lasso)
colorcode <- rep("grey",N)
colorcode[c(45,53)] <- "deeppink1"
par(mfrow=c(3, 3))
for (i in 1:N) {
if (min(features_lasso[,i])>0){
hist(log(features_lasso[,i]), breaks = 20, main = paste(i,paste("log(",names(features_lasso)[i],")",sep = ""),sep = ".\n"), border=colorcode[i], col=colorcode[i],cex.main=0.9)
}
}
Generalization of Error
Once again, we are going to assess the performance of Model 2.2 (with Lasso Regression). Then, we are going to assume the test errors as an approximation of our generalization error.
# predict on test set
pred_test_lasso2.2 <- predict(lasso2.2, s=lasso2.2$lambda.1se, newx=xmat_test)
# MSE
rmse_test2.2 <- RMSE(pred_test_lasso2.2, test_set$critical_temp)
cat("\nLinear Regression with Regularisation: test MSE = ",rmse_test2.2)##
## Linear Regression with Regularisation: test MSE = 16.58278# R-squared
rsq_test2.2 <- cor(pred_test_lasso2.2, test_set$critical_temp)^2
cat("\nLinear Regression with Regularisation: test adj.rsquared = ",rsq_test2.2)##
## Linear Regression with Regularisation: test adj.rsquared = 0.7536771Visualization of Generalized Error
The visualization on the test results once agian tell the same story. The weak relationship between X values and Y values appeared within the interval of Y < 100.
# Visualizing performance/error on test set
Lasso_regression_test <- ggplot() +
geom_point(aes(x = test_set$critical_temp, y = pred_test_lasso2.2),
colour = 'magenta',alpha=0.5,size=3) +
ggtitle('Lasso Regression') +
ylab('Prediction') +
xlab('True Value') +
theme_minimal() +
geom_abline(colour = "grey80", size = 1)
Lasso_regression_test 
3.1.5 Linear Regression with Feature Crosses
We suspect that the relationship between features and the label was a non-linear one as we saw that the linear regression trainined on the raw features and even the log-transformed ones did poorly on the test set. Here we try to solve a nonlinear problem through feature enginerring to generate feature crosses.
A feature cross is a synthetic feature that encodes nonlinearity in the feature space by multiplying two or more input features together. (The term cross comes from cross product).
Let’s try brute force all the possible combinations of two-way crosses all at once: \(x_3 = x_{1} x_2\) (Although we may suffer from heavy complexity of the model due to a significant increas in the number of features and at at the risk of overfitting, we still wanted to give it a shot!)
Also, due to the skewness we observed in the distribution of critical_temp, our target variable, we are going to take the log of it. Let’s try and see if it helps improve the prediction accuracy.
par(mfrow=c(2, 2))
hist(superconductor$critical_temp, breaks = 25, main = "Tc", border="grey", col="dimgrey")
hist(log(superconductor$critical_temp), breaks = 25, main = "log(Tc)", border="grey", col="dimgrey")
3.1.5.1 Linear Regression on two-way feature crosses
fit data to the linear regression model.
fit3.1 = lm(log(critical_temp)~(.)^2,data = training_set)This time, we got a very high R-squared of a little over 0.9 using up to 3321 predictors. Although it is a good number to see here, we might still wonder if it is the result of over-fitting given to the complexity.
We will check how good our test data fits the model to decide that this is really a useful model.
num_features3.1 <- dim(summary(fit3.1)$coefficients)[1]-1
cat("Number of features in model 3.1 = ",num_features3.1)## Number of features in model 3.1 = 3321cat("\nRsquared = ",summary(fit3.1)$adj.r.squared)##
## Rsquared = 0.9066994Diagnostic plots
We run a residual analysis by plotting the diagnostic plots and try to understand what is going on with the residuals.
- Residuals vs Fitted : shows the residuals are almost evenly distributed around zero, this suggests that the model doesn’t capture the non-linear relationship
- Normal Q-Q plot shows that the residuals do not follow the straight line well, indicating non-normal distribution
- Scale-Location plot tells us the residuals are not randomly spread, violating the assumption of equal variance (homoscedasticity)
- Residuals vs Leverage shows little evididence of influencial cases
autoplot(fit3.1)
RMSE of 2-way and 3-way Feature Crosses
Let’s Predict Tc for training set and validation set and calculated the RMSE
## MODEL 3.1
# Predicting Tc for training/validation set
predLog_train3.1 <- predict(fit3.1, newdata = training_set)
pred_train3.1 <- exp(predLog_train3.1)
predLog_v3.1 <- predict(fit3.1, newdata = validation_set)
pred_v3.1 <- exp(predLog_v3.1)
# Calculates RMSE of training pred
rmse_train3.1 <- RMSE(pred_train3.1,training_set$critical_temp)
rmse_v3.1 <- RMSE(pred_v3.1, validation_set$critical_temp)
## MODEL 3.2
# Predicting Tc for training/validation set
predLog_train3.2 <- predict(fit3.2, newdata = training_set)
pred_train3.2 <- exp(predLog_train3.2)
predLog_v3.2 <- predict(fit3.2, newdata = validation_set)
pred_v3.2 <- exp(predLog_v3.2)
# Calculates RMSE of training pred
rmse_train3.2 <- RMSE(pred_train3.2,training_set$critical_temp)
rmse_v3.2 <- RMSE(pred_v3.2, validation_set$critical_temp)Assessment on Validation Set
From the table we put together below, we found something interesting.
The MSE of predictions of model 3.1 and model 3.2 on the training set was only a little over 12, while it became an incredibly huge number for the validation set. So there’s clearly a sign of overfitting, where we got a low mse on the training set while an extremely bad one on the validation set.
But is it hopeless?
Did the model do a really terrible job predicting Tc in general? Or was there just a few extreme predictions that caused such crazy MSE? Let’s find out by simply plotting the results.
lin_reg_featureCrosses2 <- c("num_predictors"=num_features3.1,
"adj.rsquared_train"=round(summary(fit3.1)$adj.r.squared,4),
"rmse_train"=round(rmse_train3.1,4),
"rmse_validation"="huge")
lin_reg_featureCrosses3 <- c("num_predictors"=num_features3.2,
"adj.rsquared_train"=round(summary(fit3.2)$adj.r.squared,4),
"rmse_train"=round(rmse_train3.2,4),
"rmse_validation"="incredibly huge")
models.featureCrosses <- data.frame(lin_reg_featureCrosses2,lin_reg_featureCrosses3)
models.featureCrosses## lin_reg_featureCrosses2 lin_reg_featureCrosses3
## num_predictors 3321 4089
## adj.rsquared_train 0.9067 0.9164
## rmse_train.1 12.571 12.0242
## rmse_validation huge incredibly hugeVisualization of Error on Training Set
There is a strong relationship going on between the true values and the predicted values. That actually explains why we got such a low mse on the training set.
# Visualizing the predicted values
ggplot() +
geom_point(aes(x = training_set$critical_temp, y = pred_train3.1),
colour = 'orangered3',alpha=0.5,size=3) +
ggtitle('Linear regression model with two-way feature crosses on trainig set') +
ylab('Prediction') +
xlab('True Value (Tc)') +
theme_minimal() +
geom_abline(colour = "grey80", size = 1)
Visualization of Error on Validation Set
Let’s move on to the test set results.
The plot was bizzare at first glance!
Then as we looked at the scale of y-axis, we soon realized that there seemed to be a few wildly extreme predictions on the validation set.
# Visualizing the predicted values excluding extreme predictions
ggplot() +
geom_point(aes(x = validation_set$critical_temp, y = pred_v3.1),
colour = 'orangered3',alpha=0.5,size=3) +
ggtitle('Linear regression model with two-way feature crosses on validation set') +
ylab('Prediction') +
xlab('True Value (Tc)') +
theme_minimal() +
geom_abline(colour = "grey80", size = 1)
As discussed, we suspected there might be some extreme predictions on the validation set.
So we filter and count the number of predicted values that were greater than 134, the maximum value in the validation set.
It turned out that there were 24 extreme values there in the predictions, which account for around 1% of the total predictions. Now let’s filter them out before plotting, otherwise the scale of y-axis would just shift significanly.
high <- max(validation_set$critical_temp)
filtered3.1 <- pred_v3.1[pred_v3.1 < high]
num_extremes3.1 <- length(pred_v3.1) - length(filtered3.1)
cat("Linear regression model with Feature Crosses: Number of extreme predictions on the validation set = ",num_extremes3.1)## Linear regression model with Feature Crosses: Number of extreme predictions on the validation set = 22cat("\nLinear regression model with Feature Crosses: Proportion of extreme predictions in the validation set =",num_extremes3.1/nrow(validation_set))##
## Linear regression model with Feature Crosses: Proportion of extreme predictions in the validation set = 0.0142487This time, we got a pretty neat plot. There’s clearly a strong positive corelation between the true values and the predicted ones. As an accurate prediction has an Y=X relationship, we are not too far from that here. We clearly did a better job with this mode than we did with the previous ones.
# Visualizing the predicted values excluding extreme predictions
ggplot() +
geom_point(aes(x = validation_set$critical_temp[pred_v3.1 < high], y = filtered3.1),
colour = 'orangered3',alpha=0.5,size=4) +
ggtitle(paste0('Linear regression model with two-way feature crosses on validation set \n(with ',num_extremes3.1,' extreme predictions removed)')) +
ylab('Prediction') +
xlab('True Value (Tc)') +
theme_minimal() +
geom_abline(colour = "grey80", size = 1)
Visualization of Error on Training Set
Now we’ll plot the preformance for model 3.2, which has more than 4000 predictors. Again, there’s a strong linear relationship between X and Y as we fitted the training set.
# Visualizing the predicted values
ggplot() +
geom_point(aes(x = training_set$critical_temp, y = pred_train3.2),
colour = 'rosybrown',alpha=0.5,size=3) +
ggtitle('Linear regression model with three-way feature crosses on training set') +
ylab('Prediction') +
xlab('True Value (Tc)') +
theme_minimal() +
geom_abline(colour = "grey80", size = 1)
Visualization of Error on Validation Set
As expected, we got a few of the crazy predictons again. Let’s find out how many of the predictions are overly extreme this time.
# Visualizing the predicted values excluding extreme predictions
ggplot() +
geom_point(aes(x = validation_set$critical_temp, y = pred_v3.2),
colour = 'rosybrown',alpha=0.5,size=3) +
ggtitle('Linear regression model with three-way feature crosses on validation set') +
ylab('Prediction') +
xlab('True Value (Tc)') +
theme_minimal() +
geom_abline(colour = "grey80", size = 1)
Turned out there are 43 predicted values > max(test_set.Tc) That was more than double the number we have observed in our model 3.1. Similarly, let’s filter them out and plot again to see the overall performance of model 3.2
filtered3.2 <- pred_v3.2[pred_v3.2 < high]
num_extremes3.2 <- length(pred_v3.2) - length(filtered3.2)
cat("Linear Model with Feature Crosses (3 way): Number of extreme predictions on the validation set = ",num_extremes3.2)## Linear Model with Feature Crosses (3 way): Number of extreme predictions on the validation set = 29cat("\nLinear Model with Feature Crosses (3 way): Proportion of extreme predictions in the validation set =",num_extremes3.2/nrow(validation_set))##
## Linear Model with Feature Crosses (3 way): Proportion of extreme predictions in the validation set = 0.01878238After removing those extreme cases, we got a much nicer plot again. The predictions had a strong linear relationship with the true Tc.
# Visualizing the predicted values excluding extreme predictions
ggplot() +
geom_point(aes(x = validation_set$critical_temp[pred_v3.2 < high], y = filtered3.2),
colour = 'rosybrown',alpha=0.5,size=4) +
ggtitle(paste0('Linear Model with Feature Crosses (3 way) Performance on Validation Set \n(with ',num_extremes3.2,' extreme predictions removed)')) +
ylab('Prediction') +
xlab('True Value (Tc)') +
theme_minimal() +
geom_abline(colour = "grey80", size = 1)
Comparing 2-way and 3-way Feature Crosses
Looking at the performance of the two models, we still believe that Model3.1 is a better choice given the number of feature, number of extreme predictions, and overall accuracy in predicting Tc.
Even though clearly both of the models are an example of overfitting, due to time and complexity, we are not going to do feature selection such as stepwise algorithm or regularization as we did before. We decided to leave it as it. Still, we will observe and analyze its performance when we try fitting the test set on it.
# Visualizing Model 3.1 error
p1<- ggplot() +
geom_point(aes(x = training_set$critical_temp, y = pred_train3.1),colour = 'orangered3',alpha=0.5,size=2) +
ggtitle('Model 3.1 Performance on Trainig Set') +
theme(plot.title = element_text(size = 10)) +
ylab('Prediction') +
xlab('True Value (Tc)') + theme_light()
# Visualizing Model 3.2 error
p2<- ggplot() +
geom_point(aes(x = training_set$critical_temp, y = pred_train3.2),colour = 'rosybrown',alpha=0.5,size=2) +
ggtitle('Model 3.2 Performance on Trainig Set') +
theme(plot.title = element_text(size = 10)) +
ylab('Prediction') +
xlab('True Value (Tc)') + theme_light()
# Visualizing Model 3.1 error after removing wild predictions
p3<- ggplot() +
geom_point(aes(x = validation_set$critical_temp[pred_v3.1 < high], y = filtered3.1),colour = 'orangered3',alpha=0.5,size=3) +
ggtitle(paste0('Model 3.1 Performance on Validation Set \n(with ',num_extremes3.1,' extreme predictions removed)')) +
theme(plot.title = element_text(size = 10)) +
ylab('Prediction') +
xlab('True Value (Tc)') + theme_light()
# Visualizing Model 3.2 results after removing wild predictions
p4<- ggplot() +
geom_point(aes(x = validation_set$critical_temp[pred_v3.2 < high], y = filtered3.2),colour = 'rosybrown',alpha=0.5,size=3) +
ggtitle(paste0('Model 3.2 Performance on Validation Set \n(with ',num_extremes3.2,' extreme predictions removed)')) +
theme(plot.title = element_text(size = 10)) +
ylab('Prediction') +
xlab('True Value (Tc)') + theme_light()
grid.arrange(p1, p2, p3,p4,nrow=2)
Generalization of Error
Now that we have chosen our third model,we are going to assess the performance of our model on the test set as usual. We already expected to see a significantly large MSE due to extreme predictions we observed in the training set and the validation set results.
# Predicting Tc on test set
predLog_test3.1 <- predict(fit3.1, newdata = test_set)
pred_test3.1 <- exp(predLog_test3.1)
# Calculates RMSE
rmse_test3.1 <- RMSE(pred_test3.1, test_set$critical_temp)
cat("\nMODEL 3.1: RMSE for the test predictions =", rmse_test3.1)##
## MODEL 3.1: RMSE for the test predictions = 758819.2Visualization of Genralized Error
Judging by the MSE, it is obvious that there are extreme predictions for sure. Turns out its less than 1%, not too bad. Let’s filter them out before plotting.
#high <- max(validation_set$critical_temp)
filtered <- pred_test3.1[pred_test3.1 < high]
num_extremes <- length(pred_test3.1) - length(filtered)
# rsquared and mse after removing extremes
rsq_test3.1 <- cor(filtered, test_set$critical_temp[pred_test3.1 < high])^2
adj.rmse_test3.1 <- RMSE(filtered,test_set$critical_temp[pred_test3.1 < high])
cat("MODEL3.1: Number of extreme predictions on the test set = ",num_extremes)## MODEL3.1: Number of extreme predictions on the test set = 22cat("\nMODEL3.1: Proportion of extreme predictions in the test set =",num_extremes/nrow(test_set))##
## MODEL3.1: Proportion of extreme predictions in the test set = 0.0142487cat("\n(After removing extreme observations) test RMSE =",adj.rmse_test3.1)##
## (After removing extreme observations) test RMSE = 13.98703#cat("\n(After removing extreme observations) adj.rsquared =",rsq_test3.1)We are going to fit the test set to our extremely complex third model and visualize the prediction. From the plot below, we can see a strong linear relationship between true values and the predicted values. Although we got around 1% extreme predictions on the test set, we still did a fairly good job on predicting Tc using our third model.
# Visualizing generalization error
featureCrosses_test <- ggplot() +
geom_point(aes(x = test_set$critical_temp[pred_test3.1 < high], y = filtered),
colour = 'orangered3',alpha=0.5,size=3) +
ggtitle(paste0('Feature Crosses (with ',num_extremes,' extreme predictions removed)')) +
ylab('Prediction') +
xlab('True Value') +
theme_minimal() +
geom_abline(colour = "grey80", size = 1)
featureCrosses_test 
3.2 Random Forest
Random Forest algorithm is an ensemble learning method for classification as well as regression that leverages the power of multiple decision trees for making decisions. Not only is this algorithm better than linear regression at capturing non-linear relationships, by introducing a technique called bagging to the traditional Decision Tree Algorithm, it also improves the problem of overfitting to the training data.
Bagging (Bootstrap aggregating) here is a method designed to reduces variance and helps to avoid overfitting. By sampling from a training set \(D\) uniformly(randomly) and with replacement, we get \(n\) different bootstrap samples for building \(m\) decision trees. We then combine the results of the \(m\) tress by averaging the output (for regression) or voting (for classification).
random_forest_model <- train(
critical_temp ~., data = training_set,
trControl = trainControl( method = "cv", number = 10, search = "random"), tuneLength = 5,
method = "ranger",
importance = 'impurity',
preProc = c("range")
)## Growing trees.. Progress: 47%. Estimated remaining time: 34 seconds.
## Growing trees.. Progress: 96%. Estimated remaining time: 2 seconds.
## Growing trees.. Progress: 44%. Estimated remaining time: 39 seconds.
## Growing trees.. Progress: 89%. Estimated remaining time: 7 seconds.
## Growing trees.. Progress: 41%. Estimated remaining time: 44 seconds.
## Growing trees.. Progress: 80%. Estimated remaining time: 15 seconds.
## Growing trees.. Progress: 66%. Estimated remaining time: 15 seconds.
## Growing trees.. Progress: 51%. Estimated remaining time: 29 seconds.
## Growing trees.. Progress: 45%. Estimated remaining time: 37 seconds.
## Growing trees.. Progress: 90%. Estimated remaining time: 7 seconds.
## Growing trees.. Progress: 42%. Estimated remaining time: 43 seconds.
## Growing trees.. Progress: 84%. Estimated remaining time: 11 seconds.
## Growing trees.. Progress: 68%. Estimated remaining time: 14 seconds.
## Growing trees.. Progress: 50%. Estimated remaining time: 31 seconds.
## Growing trees.. Progress: 47%. Estimated remaining time: 35 seconds.
## Growing trees.. Progress: 91%. Estimated remaining time: 6 seconds.
## Growing trees.. Progress: 41%. Estimated remaining time: 44 seconds.
## Growing trees.. Progress: 82%. Estimated remaining time: 13 seconds.
## Growing trees.. Progress: 69%. Estimated remaining time: 13 seconds.
## Growing trees.. Progress: 46%. Estimated remaining time: 35 seconds.
## Growing trees.. Progress: 96%. Estimated remaining time: 2 seconds.
## Growing trees.. Progress: 44%. Estimated remaining time: 38 seconds.
## Growing trees.. Progress: 91%. Estimated remaining time: 5 seconds.
## Growing trees.. Progress: 31%. Estimated remaining time: 1 minute, 7 seconds.
## Growing trees.. Progress: 52%. Estimated remaining time: 58 seconds.
## Growing trees.. Progress: 87%. Estimated remaining time: 13 seconds.
## Growing trees.. Progress: 61%. Estimated remaining time: 19 seconds.
## Growing trees.. Progress: 51%. Estimated remaining time: 29 seconds.
## Growing trees.. Progress: 43%. Estimated remaining time: 41 seconds.
## Growing trees.. Progress: 87%. Estimated remaining time: 8 seconds.
## Growing trees.. Progress: 41%. Estimated remaining time: 44 seconds.
## Growing trees.. Progress: 83%. Estimated remaining time: 13 seconds.
## Growing trees.. Progress: 65%. Estimated remaining time: 16 seconds.
## Growing trees.. Progress: 48%. Estimated remaining time: 33 seconds.
## Growing trees.. Progress: 99%. Estimated remaining time: 0 seconds.
## Growing trees.. Progress: 45%. Estimated remaining time: 37 seconds.
## Growing trees.. Progress: 90%. Estimated remaining time: 6 seconds.
## Growing trees.. Progress: 40%. Estimated remaining time: 46 seconds.
## Growing trees.. Progress: 67%. Estimated remaining time: 30 seconds.
## Growing trees.. Progress: 62%. Estimated remaining time: 19 seconds.
## Growing trees.. Progress: 49%. Estimated remaining time: 32 seconds.
## Growing trees.. Progress: 99%. Estimated remaining time: 0 seconds.
## Growing trees.. Progress: 43%. Estimated remaining time: 41 seconds.
## Growing trees.. Progress: 85%. Estimated remaining time: 11 seconds.
## Growing trees.. Progress: 34%. Estimated remaining time: 1 minute, 0 seconds.
## Growing trees.. Progress: 70%. Estimated remaining time: 26 seconds.
## Growing trees.. Progress: 61%. Estimated remaining time: 20 seconds.
## Growing trees.. Progress: 47%. Estimated remaining time: 34 seconds.
## Growing trees.. Progress: 96%. Estimated remaining time: 2 seconds.
## Growing trees.. Progress: 44%. Estimated remaining time: 39 seconds.
## Growing trees.. Progress: 89%. Estimated remaining time: 7 seconds.
## Growing trees.. Progress: 39%. Estimated remaining time: 48 seconds.
## Growing trees.. Progress: 78%. Estimated remaining time: 17 seconds.
## Growing trees.. Progress: 60%. Estimated remaining time: 21 seconds.
## Growing trees.. Progress: 49%. Estimated remaining time: 31 seconds.
## Growing trees.. Progress: 98%. Estimated remaining time: 1 seconds.
## Growing trees.. Progress: 46%. Estimated remaining time: 36 seconds.
## Growing trees.. Progress: 91%. Estimated remaining time: 5 seconds.
## Growing trees.. Progress: 41%. Estimated remaining time: 44 seconds.
## Growing trees.. Progress: 82%. Estimated remaining time: 13 seconds.
## Growing trees.. Progress: 60%. Estimated remaining time: 20 seconds.
## Growing trees.. Progress: 48%. Estimated remaining time: 33 seconds.
## Growing trees.. Progress: 98%. Estimated remaining time: 1 seconds.
## Growing trees.. Progress: 47%. Estimated remaining time: 35 seconds.
## Growing trees.. Progress: 85%. Estimated remaining time: 11 seconds.
## Growing trees.. Progress: 38%. Estimated remaining time: 51 seconds.
## Growing trees.. Progress: 78%. Estimated remaining time: 17 seconds.
## Growing trees.. Progress: 65%. Estimated remaining time: 16 seconds.
## Growing trees.. Progress: 59%. Estimated remaining time: 21 seconds.# random forest training results
random_forest_model$results## min.node.size mtry splitrule RMSE Rsquared MAE RMSESD
## 4 13 10 extratrees 9.677616 0.9213217 5.682707 0.4860627
## 1 2 16 maxstat 9.446636 0.9247586 5.388818 0.4859528
## 2 6 24 maxstat 9.552107 0.9231087 5.467311 0.4914655
## 3 8 28 maxstat 9.611306 0.9222023 5.514149 0.5018636
## 5 19 35 maxstat 9.990829 0.9163405 5.853876 0.5169983
## RsquaredSD MAESD
## 4 0.007411564 0.1870933
## 1 0.007221282 0.1571687
## 2 0.007378714 0.1528034
## 3 0.007588542 0.1506292
## 5 0.008185058 0.1570023# random forest with optimised tuning perameters
random_forest_model$bestTune## mtry splitrule min.node.size
## 1 16 maxstat 2random_forest_model$finalModel## Ranger result
##
## Call:
## ranger::ranger(dependent.variable.name = ".outcome", data = x, mtry = min(param$mtry, ncol(x)), min.node.size = param$min.node.size, splitrule = as.character(param$splitrule), write.forest = TRUE, probability = classProbs, ...)
##
## Type: Regression
## Number of trees: 500
## Sample size: 17290
## Number of independent variables: 81
## Mtry: 16
## Target node size: 2
## Variable importance mode: impurity
## Splitrule: maxstat
## OOB prediction error (MSE): 86.47003
## R squared (OOB): 0.9268189# List of features with their importance scores
rf_features <- varImp(random_forest_model)
print(rf_features$importance)## Overall
## number_of_elements 8.6845292
## mean_atomic_mass 15.8868933
## wtd_mean_atomic_mass 59.1544978
## gmean_atomic_mass 13.3642900
## wtd_gmean_atomic_mass 56.5833917
## entropy_atomic_mass 1.1004045
## wtd_entropy_atomic_mass 63.1820627
## range_atomic_mass 28.3385286
## wtd_range_atomic_mass 70.8850100
## std_atomic_mass 13.6316424
## wtd_std_atomic_mass 64.6598559
## mean_fie 14.8357405
## wtd_mean_fie 58.6167339
## gmean_fie 17.1887229
## wtd_gmean_fie 57.6471647
## entropy_fie 0.2422649
## wtd_entropy_fie 55.1561265
## range_fie 19.8088822
## wtd_range_fie 70.1847525
## std_fie 3.8029914
## wtd_std_fie 67.6328522
## mean_atomic_radius 28.8765014
## wtd_mean_atomic_radius 66.6681244
## gmean_atomic_radius 13.3784045
## wtd_gmean_atomic_radius 60.6103713
## entropy_atomic_radius 3.1937353
## wtd_entropy_atomic_radius 62.4846252
## range_atomic_radius 26.7373524
## wtd_range_atomic_radius 72.3394854
## std_atomic_radius 12.2705662
## wtd_std_atomic_radius 70.8406127
## mean_Density 8.8726992
## wtd_mean_Density 57.7426850
## gmean_Density 4.9821414
## wtd_gmean_Density 54.2331090
## entropy_Density 4.0194418
## wtd_entropy_Density 71.2375102
## range_Density 28.7959942
## wtd_range_Density 71.9234605
## std_Density 13.4396785
## wtd_std_Density 65.8824058
## mean_ElectronAffinity 23.8004288
## wtd_mean_ElectronAffinity 71.5723351
## gmean_ElectronAffinity 24.2157853
## wtd_gmean_ElectronAffinity 78.7209820
## entropy_ElectronAffinity 11.4713598
## wtd_entropy_ElectronAffinity 69.2795560
## range_ElectronAffinity 31.3870785
## wtd_range_ElectronAffinity 79.0258032
## std_ElectronAffinity 15.1933557
## wtd_std_ElectronAffinity 75.1397141
## mean_FusionHeat 8.2237335
## wtd_mean_FusionHeat 63.0281014
## gmean_FusionHeat 6.4177293
## wtd_gmean_FusionHeat 59.3875950
## entropy_FusionHeat 2.3745707
## wtd_entropy_FusionHeat 61.1154316
## range_FusionHeat 29.8008345
## wtd_range_FusionHeat 69.6382108
## std_FusionHeat 14.2143830
## wtd_std_FusionHeat 73.0560909
## mean_ThermalConductivity 17.3026121
## wtd_mean_ThermalConductivity 76.4047757
## gmean_ThermalConductivity 11.1037361
## wtd_gmean_ThermalConductivity 71.1468012
## entropy_ThermalConductivity 15.9099027
## wtd_entropy_ThermalConductivity 79.0912514
## range_ThermalConductivity 0.0000000
## wtd_range_ThermalConductivity 81.1859772
## std_ThermalConductivity 10.1293290
## wtd_std_ThermalConductivity 75.2437346
## mean_Valence 40.1334009
## wtd_mean_Valence 100.0000000
## gmean_Valence 27.4427536
## wtd_gmean_Valence 90.7907407
## entropy_Valence 24.4838578
## wtd_entropy_Valence 75.3036692
## range_Valence 17.6967426
## wtd_range_Valence 89.4173170
## std_Valence 42.9398988
## wtd_std_Valence 91.0774620Generalization of Error
# Predicting Tc for training/test
pred_test_rf = predict(random_forest_model, newdata = test_set)
rmse_test_rf <- RMSE(pred_test_rf,test_set$critical_temp)
# Calculates RMSE of training pred
cat("\nRandom Forest Model: RMSE for the test predictions =", rmse_test_rf)##
## Random Forest Model: RMSE for the test predictions = 10.21365Visualization of Gernerailzed Error
# Visualizing the fit
rf_test <- ggplot() +
geom_point(aes(x = test_set$critical_temp, y = pred_test_rf),
colour = 'gold',alpha=0.5,size=3) +
ggtitle('Random Forest') +
ylab('Prediction') +
xlab('True Value (Tc)') +
theme_minimal() +
geom_abline(colour = "grey80", size = 1)
rf_test
3.3 Gradient Boosting
In addition to Random Forest, we will try another ensemble learning method called Gradient Boosting. Similar to Random Firest algorithm, it builds a set of decision tress with bootstrapping and then combine the results to make a prediction.
However, the way they build those trees are different. Random Forests build each tree independently, thus it doesnt matter in what order you build the tress. On the other hand, Gradient Boosting algorithm builds one tree at a time, introducing a new tree to improve the mistakes (residuals) made by the previous one by trying to minimise the loss using gradient descent algorithm. Similar to other boosting methods, gradient boosting combines weak “learners” into a single strong learner in such iterative fashion.
# Gradient Boosting Machine
gb_model <- train(
critical_temp ~ ., data = training_set,
method = "gbm",
tuneLength = 3,
preProc = c("range"),
trControl = trainControl( method = "cv", number = 10, search = "random", verbose = FALSE)
) ## Iter TrainDeviance ValidDeviance StepSize Improve
## 1 820.2913 nan 0.2306 360.3121
## 2 598.6821 nan 0.2306 219.2438
## 3 458.8542 nan 0.2306 138.1983
## 4 373.7684 nan 0.2306 83.2345
## 5 315.4238 nan 0.2306 54.8164
## 6 277.8400 nan 0.2306 36.3623
## 7 251.1454 nan 0.2306 25.3507
## 8 232.6973 nan 0.2306 18.0249
## 9 219.7588 nan 0.2306 12.2491
## 10 208.3323 nan 0.2306 11.0855
## 20 163.8643 nan 0.2306 1.0526
## 40 131.4026 nan 0.2306 0.5118
## 60 115.1072 nan 0.2306 0.3414
## 80 105.9440 nan 0.2306 0.0299
## 100 97.7523 nan 0.2306 -0.0345
## 120 89.6829 nan 0.2306 0.0948
## 140 83.3906 nan 0.2306 0.0341
## 160 78.3659 nan 0.2306 -0.2009
## 180 74.8566 nan 0.2306 -0.0881
## 200 71.1983 nan 0.2306 -0.2661
## 220 67.7012 nan 0.2306 -0.1934
## 240 64.7260 nan 0.2306 -0.2320
## 260 62.2020 nan 0.2306 -0.1965
## 280 60.0036 nan 0.2306 -0.0816
## 300 58.2190 nan 0.2306 -0.1936
## 320 55.9247 nan 0.2306 -0.0902
## 340 54.3282 nan 0.2306 -0.1183
## 360 52.7059 nan 0.2306 -0.1692
## 380 51.3612 nan 0.2306 -0.1564
## 400 50.0680 nan 0.2306 -0.2155
## 420 48.8575 nan 0.2306 -0.1558
## 440 47.6979 nan 0.2306 -0.0837
## 460 46.8128 nan 0.2306 -0.1121
## 480 45.9042 nan 0.2306 -0.1624
## 500 44.9538 nan 0.2306 -0.1455
## 520 44.0979 nan 0.2306 -0.1494
## 540 43.3918 nan 0.2306 -0.1428
## 560 42.7064 nan 0.2306 -0.1303
## 580 41.9847 nan 0.2306 -0.0993
## 600 41.1431 nan 0.2306 -0.1838
## 620 40.5524 nan 0.2306 -0.1410
## 640 39.9493 nan 0.2306 -0.0605
## 660 39.2882 nan 0.2306 -0.0877
## 680 38.7405 nan 0.2306 -0.1226
## 700 38.2163 nan 0.2306 -0.1081
## 720 37.6423 nan 0.2306 -0.1124
## 740 37.1198 nan 0.2306 -0.1088
## 760 36.7274 nan 0.2306 -0.1317
## 780 36.3968 nan 0.2306 -0.0921
## 800 36.0118 nan 0.2306 -0.1603
## 820 35.5982 nan 0.2306 -0.1107
## 840 35.2645 nan 0.2306 -0.0910
## 860 34.8935 nan 0.2306 -0.1514
## 880 34.5733 nan 0.2306 -0.1917
## 900 34.3070 nan 0.2306 -0.1004
## 920 33.9854 nan 0.2306 -0.1446
## 940 33.6604 nan 0.2306 -0.1147
## 960 33.4058 nan 0.2306 -0.1391
## 980 33.0602 nan 0.2306 -0.0867
## 1000 32.8290 nan 0.2306 -0.1165
## 1020 32.5714 nan 0.2306 -0.1629
## 1040 32.3317 nan 0.2306 -0.1460
## 1060 32.1684 nan 0.2306 -0.0894
## 1080 31.9288 nan 0.2306 -0.0630
## 1100 31.7255 nan 0.2306 -0.1002
## 1120 31.5634 nan 0.2306 -0.1715
## 1140 31.3720 nan 0.2306 -0.1794
## 1160 31.0873 nan 0.2306 -0.0874
## 1180 30.9373 nan 0.2306 -0.1035
## 1200 30.7372 nan 0.2306 -0.2175
## 1220 30.4989 nan 0.2306 -0.0935
## 1240 30.3589 nan 0.2306 -0.0984
## 1260 30.1654 nan 0.2306 -0.0604
## 1280 29.9929 nan 0.2306 -0.1216
## 1300 29.7950 nan 0.2306 -0.1415
## 1320 29.6314 nan 0.2306 -0.0903
## 1340 29.4563 nan 0.2306 -0.1152
## 1360 29.2934 nan 0.2306 -0.1568
## 1380 29.1701 nan 0.2306 -0.1029
## 1400 29.0160 nan 0.2306 -0.1165
## 1420 28.8555 nan 0.2306 -0.1610
## 1440 28.7316 nan 0.2306 -0.1167
## 1460 28.6386 nan 0.2306 -0.0900
## 1480 28.5134 nan 0.2306 -0.1345
## 1500 28.4165 nan 0.2306 -0.2115
## 1520 28.2953 nan 0.2306 -0.2164
## 1540 28.1938 nan 0.2306 -0.1576
## 1560 28.0855 nan 0.2306 -0.2179
## 1580 27.9710 nan 0.2306 -0.1375
## 1600 27.8689 nan 0.2306 -0.1397
## 1620 27.7130 nan 0.2306 -0.1156
## 1640 27.6342 nan 0.2306 -0.1429
## 1660 27.5094 nan 0.2306 -0.1010
## 1680 27.5106 nan 0.2306 -0.1106
## 1700 27.3978 nan 0.2306 -0.1043
## 1720 27.2967 nan 0.2306 -0.0367
## 1740 27.1649 nan 0.2306 -0.1097
## 1760 27.0857 nan 0.2306 -0.1493
## 1780 26.9950 nan 0.2306 -0.1102
## 1800 26.9209 nan 0.2306 -0.0597
## 1820 26.8076 nan 0.2306 -0.1342
## 1840 26.7148 nan 0.2306 -0.1491
## 1860 26.6261 nan 0.2306 -0.1361
## 1880 26.5313 nan 0.2306 -0.0910
## 1900 26.4739 nan 0.2306 -0.1533
## 1920 26.3820 nan 0.2306 -0.1307
## 1940 26.2836 nan 0.2306 -0.0640
## 1960 26.2352 nan 0.2306 -0.1835
## 1980 26.1312 nan 0.2306 -0.1352
## 2000 26.0805 nan 0.2306 -0.1031
## 2020 26.0557 nan 0.2306 -0.1113
## 2040 25.9696 nan 0.2306 -0.0860
## 2060 25.8802 nan 0.2306 -0.1068
## 2080 25.8650 nan 0.2306 -0.0787
## 2100 25.7584 nan 0.2306 -0.1191
## 2120 25.7463 nan 0.2306 -0.0644
## 2140 25.6889 nan 0.2306 -0.1365
## 2160 25.6601 nan 0.2306 -0.1187
## 2180 25.6439 nan 0.2306 -0.1453
## 2200 25.5546 nan 0.2306 -0.0447
## 2220 25.5277 nan 0.2306 -0.1155
## 2240 25.4004 nan 0.2306 -0.1281
## 2260 25.3826 nan 0.2306 -0.1152
## 2280 25.3321 nan 0.2306 -0.1384
## 2300 25.2527 nan 0.2306 -0.1650
## 2320 25.2276 nan 0.2306 -0.1108
## 2340 25.2350 nan 0.2306 -0.2275
## 2360 25.1238 nan 0.2306 -0.0568
## 2380 25.0739 nan 0.2306 -0.1109
## 2400 25.1065 nan 0.2306 -0.1050
## 2420 25.0529 nan 0.2306 -0.1090
## 2440 25.0301 nan 0.2306 -0.1336
## 2460 24.9988 nan 0.2306 -0.0874
## 2480 24.9461 nan 0.2306 -0.1137
## 2500 24.8900 nan 0.2306 -0.1126
## 2520 24.8626 nan 0.2306 -0.2967
## 2540 24.8382 nan 0.2306 -0.1438
## 2560 24.7570 nan 0.2306 -0.1385
## 2580 24.7368 nan 0.2306 -0.1311
## 2600 24.6795 nan 0.2306 -0.1092
## 2620 24.6338 nan 0.2306 -0.1396
## 2640 24.5671 nan 0.2306 -0.1368
## 2660 24.5834 nan 0.2306 -0.3166
## 2680 24.5194 nan 0.2306 -0.1249
## 2700 24.5075 nan 0.2306 -0.1874
## 2720 24.4778 nan 0.2306 -0.1905
## 2740 24.4080 nan 0.2306 -0.1227
## 2760 24.3755 nan 0.2306 -0.1121
## 2780 24.3471 nan 0.2306 -0.1290
## 2800 24.2629 nan 0.2306 -0.1238
## 2820 24.2342 nan 0.2306 -0.1832
## 2840 24.2029 nan 0.2306 -0.1271
## 2860 24.1922 nan 0.2306 -0.0790
## 2880 24.2217 nan 0.2306 -0.1198
## 2900 24.1293 nan 0.2306 -0.1128
## 2920 24.0930 nan 0.2306 -0.1847
## 2940 24.0430 nan 0.2306 -0.0950
## 2960 24.0120 nan 0.2306 -0.0868
## 2980 23.9303 nan 0.2306 -0.0435
## 3000 23.9434 nan 0.2306 -0.0825
## 3020 23.9519 nan 0.2306 -0.0914
## 3040 23.9029 nan 0.2306 -0.0851
## 3060 23.9317 nan 0.2306 -0.1806
## 3080 23.8765 nan 0.2306 -0.1483
## 3100 23.8262 nan 0.2306 -0.0634
## 3120 23.8491 nan 0.2306 -0.1528
## 3140 23.7744 nan 0.2306 -0.1464
## 3160 23.7189 nan 0.2306 -0.1142
## 3180 23.7124 nan 0.2306 -0.0769
## 3200 23.6925 nan 0.2306 -0.1048
## 3220 23.6352 nan 0.2306 -0.1246
## 3240 23.6419 nan 0.2306 -0.1481
## 3260 23.6003 nan 0.2306 -0.0593
## 3280 23.6261 nan 0.2306 -0.1442
## 3300 23.5481 nan 0.2306 -0.1133
## 3320 23.5229 nan 0.2306 -0.0772
## 3340 23.4981 nan 0.2306 -0.1451
## 3360 23.4716 nan 0.2306 -0.0901
## 3380 23.4419 nan 0.2306 -0.0773
## 3400 23.4127 nan 0.2306 -0.0964
## 3420 23.3991 nan 0.2306 -0.1600
## 3440 23.4046 nan 0.2306 -0.1901
## 3460 23.3581 nan 0.2306 -0.0651
## 3480 23.3325 nan 0.2306 -0.2195
## 3500 23.3037 nan 0.2306 -0.1776
## 3520 23.2473 nan 0.2306 -0.1053
## 3540 23.2146 nan 0.2306 -0.1009
## 3560 23.2291 nan 0.2306 -0.1352
## 3580 23.2069 nan 0.2306 -0.1686
## 3600 23.1637 nan 0.2306 -0.0974
## 3620 23.1529 nan 0.2306 -0.0468
## 3640 23.1079 nan 0.2306 -0.1448
## 3660 23.1358 nan 0.2306 -0.0604
## 3680 23.1074 nan 0.2306 -0.1201
## 3700 23.0796 nan 0.2306 -0.1680
## 3720 23.0987 nan 0.2306 -0.2064
##
## Iter TrainDeviance ValidDeviance StepSize Improve
## 1 671.7691 nan 0.3896 501.6199
## 2 474.4354 nan 0.3896 195.3405
## 3 377.6369 nan 0.3896 96.6095
## 4 328.6915 nan 0.3896 47.3673
## 5 295.8788 nan 0.3896 29.8144
## 6 278.0775 nan 0.3896 17.7912
## 7 264.9746 nan 0.3896 12.1795
## 8 255.9029 nan 0.3896 8.5932
## 9 248.8889 nan 0.3896 5.7788
## 10 240.9638 nan 0.3896 7.3431
## 20 206.6818 nan 0.3896 1.9613
## 40 172.4850 nan 0.3896 1.0346
## 60 154.3235 nan 0.3896 0.3021
## 80 141.8703 nan 0.3896 -0.2541
## 100 133.5588 nan 0.3896 -0.2647
## 120 126.5693 nan 0.3896 -0.0190
## 140 121.7164 nan 0.3896 -0.1568
## 160 115.3749 nan 0.3896 -0.0796
## 180 111.5530 nan 0.3896 -0.1101
## 200 108.4061 nan 0.3896 -0.0267
## 220 105.0297 nan 0.3896 -0.2647
## 240 102.0045 nan 0.3896 -0.0548
## 260 99.2862 nan 0.3896 -0.2295
## 280 96.7103 nan 0.3896 -0.0604
## 300 94.3827 nan 0.3896 -0.4061
## 320 91.6240 nan 0.3896 -0.0495
## 340 89.7554 nan 0.3896 -0.0432
## 360 87.5360 nan 0.3896 -0.1834
## 380 85.3331 nan 0.3896 -0.1846
## 400 83.7717 nan 0.3896 -0.1872
## 420 82.1428 nan 0.3896 -0.1317
## 440 80.5306 nan 0.3896 -0.1857
## 460 79.2666 nan 0.3896 -0.0648
## 480 78.0451 nan 0.3896 -0.1527
## 500 77.0711 nan 0.3896 -0.1370
## 520 75.3480 nan 0.3896 -0.0973
## 540 74.0730 nan 0.3896 -0.2305
## 560 73.0420 nan 0.3896 -0.0689
## 580 72.2410 nan 0.3896 -0.2052
## 600 71.1474 nan 0.3896 -0.1227
## 620 69.9681 nan 0.3896 -0.0758
## 640 69.1475 nan 0.3896 -0.1941
## 660 67.9939 nan 0.3896 -0.1398
## 680 67.2168 nan 0.3896 -0.1334
## 700 66.3240 nan 0.3896 -0.1215
## 720 65.6605 nan 0.3896 -0.0365
## 740 64.8266 nan 0.3896 -0.0781
## 760 63.9765 nan 0.3896 -0.1381
## 780 63.2592 nan 0.3896 -0.1404
## 800 62.5115 nan 0.3896 -0.1707
## 820 62.0700 nan 0.3896 -0.1791
## 840 61.4731 nan 0.3896 -0.0833
## 860 60.8101 nan 0.3896 -0.2099
## 880 60.3130 nan 0.3896 -0.5128
## 900 59.6949 nan 0.3896 -0.1624
## 920 59.0816 nan 0.3896 -0.0988
## 940 58.6371 nan 0.3896 -0.0170
## 960 58.1970 nan 0.3896 -0.0589
## 980 57.3731 nan 0.3896 -0.0842
## 1000 56.8640 nan 0.3896 -0.0825
## 1020 56.3813 nan 0.3896 -0.0871
## 1040 55.8749 nan 0.3896 -0.1240
## 1060 55.3825 nan 0.3896 -0.0421
## 1080 54.7306 nan 0.3896 -0.1590
## 1100 54.3654 nan 0.3896 -0.0427
## 1120 53.8275 nan 0.3896 -0.0261
## 1140 53.3908 nan 0.3896 -0.1098
## 1160 53.0258 nan 0.3896 -0.2421
## 1180 52.6596 nan 0.3896 -0.1003
## 1200 52.2419 nan 0.3896 -0.0690
## 1220 51.8385 nan 0.3896 -0.1322
## 1240 51.3682 nan 0.3896 -0.0773
## 1260 51.0852 nan 0.3896 -0.1177
## 1280 50.6896 nan 0.3896 -0.1070
## 1300 50.3952 nan 0.3896 -0.1382
## 1320 50.0191 nan 0.3896 -0.0859
## 1340 49.7160 nan 0.3896 -0.1043
## 1360 49.5169 nan 0.3896 -0.0988
## 1380 49.2134 nan 0.3896 -0.0785
## 1400 49.0756 nan 0.3896 -0.0580
## 1420 48.8089 nan 0.3896 -0.1056
## 1440 48.4608 nan 0.3896 -0.1440
## 1460 48.0499 nan 0.3896 -0.0908
## 1480 47.6970 nan 0.3896 -0.0576
## 1500 47.5189 nan 0.3896 -0.0790
## 1520 47.3030 nan 0.3896 -0.1736
## 1540 47.0231 nan 0.3896 -0.1310
## 1560 46.7769 nan 0.3896 -0.0763
## 1580 46.6143 nan 0.3896 -0.1641
## 1600 46.3225 nan 0.3896 -0.1014
## 1620 46.0970 nan 0.3896 -0.0779
## 1640 45.9173 nan 0.3896 -0.0890
## 1660 45.6981 nan 0.3896 -0.1062
## 1680 45.5261 nan 0.3896 -0.0778
## 1700 45.2996 nan 0.3896 -0.0604
## 1720 45.1532 nan 0.3896 -0.0993
## 1740 44.9579 nan 0.3896 -0.1452
## 1760 44.7753 nan 0.3896 -0.0942
## 1780 44.5757 nan 0.3896 -0.1614
## 1800 44.3033 nan 0.3896 -0.1551
## 1820 44.0760 nan 0.3896 -0.1071
## 1840 43.9670 nan 0.3896 -0.1256
## 1860 43.8818 nan 0.3896 -0.1829
## 1880 43.7356 nan 0.3896 -0.1085
## 1900 43.4918 nan 0.3896 -0.0669
## 1920 43.3152 nan 0.3896 -0.0243
## 1940 43.1474 nan 0.3896 -0.1047
## 1960 43.0721 nan 0.3896 -0.0430
## 1980 42.8295 nan 0.3896 -0.0893
## 2000 42.5775 nan 0.3896 -0.0999
## 2020 42.4204 nan 0.3896 -0.0681
## 2040 42.2921 nan 0.3896 -0.0633
## 2060 42.1689 nan 0.3896 -0.1324
## 2080 41.9684 nan 0.3896 -0.0552
## 2100 41.8009 nan 0.3896 -0.0821
## 2120 41.6404 nan 0.3896 -0.0882
## 2140 41.4944 nan 0.3896 -0.0880
## 2160 41.4587 nan 0.3896 -0.1410
## 2180 41.2788 nan 0.3896 -0.0872
## 2200 41.1513 nan 0.3896 -0.1958
## 2220 41.0049 nan 0.3896 -0.1596
## 2240 40.7971 nan 0.3896 -0.0520
## 2260 40.5702 nan 0.3896 -0.0399
## 2280 40.4627 nan 0.3896 -0.1076
## 2300 40.3633 nan 0.3896 -0.0590
## 2320 40.2796 nan 0.3896 -0.1095
## 2340 40.1879 nan 0.3896 -0.1353
## 2360 40.0313 nan 0.3896 -0.0627
## 2380 39.9332 nan 0.3896 -0.0622
## 2400 39.8443 nan 0.3896 -0.1181
## 2420 39.7230 nan 0.3896 -0.0285
## 2440 39.6195 nan 0.3896 -0.0676
## 2460 39.4445 nan 0.3896 -0.0467
## 2480 39.3484 nan 0.3896 -0.1022
## 2500 39.1918 nan 0.3896 -0.0957
## 2520 39.1115 nan 0.3896 -0.1106
## 2540 39.0168 nan 0.3896 -0.1716
## 2560 38.8884 nan 0.3896 -0.0670
## 2580 38.7715 nan 0.3896 -0.0930
## 2600 38.7386 nan 0.3896 -0.0880
## 2620 38.6067 nan 0.3896 -0.0490
## 2640 38.5340 nan 0.3896 -0.0877
## 2660 38.3115 nan 0.3896 -0.1291
## 2680 38.2716 nan 0.3896 -0.0846
## 2700 38.1455 nan 0.3896 -0.1166
## 2720 38.0508 nan 0.3896 -0.1213
## 2740 37.9538 nan 0.3896 -0.1311
## 2760 37.8647 nan 0.3896 -0.0672
## 2780 37.6963 nan 0.3896 -0.0420
## 2800 37.6276 nan 0.3896 -0.1571
## 2820 37.5078 nan 0.3896 -0.1000
## 2840 37.3660 nan 0.3896 -0.1027
## 2860 37.3271 nan 0.3896 -0.0987
## 2880 37.2108 nan 0.3896 -0.1287
## 2900 37.1023 nan 0.3896 -0.0733
## 2920 36.9750 nan 0.3896 -0.0977
## 2940 36.8505 nan 0.3896 -0.1107
## 2960 36.8358 nan 0.3896 -0.1205
## 2980 36.7271 nan 0.3896 -0.0633
## 3000 36.6034 nan 0.3896 -0.1032
## 3020 36.5134 nan 0.3896 -0.0617
## 3040 36.4105 nan 0.3896 -0.0785
## 3060 36.3294 nan 0.3896 -0.0572
## 3080 36.2050 nan 0.3896 -0.1324
## 3100 36.1944 nan 0.3896 -0.3713
## 3120 36.1019 nan 0.3896 -0.0473
## 3140 35.9728 nan 0.3896 -0.0953
## 3160 35.8641 nan 0.3896 -0.0999
## 3180 35.7981 nan 0.3896 -0.0864
## 3200 35.7123 nan 0.3896 -0.0912
## 3220 35.6955 nan 0.3896 -0.0850
## 3240 35.5785 nan 0.3896 -0.0639
## 3260 35.4647 nan 0.3896 -0.3206
## 3280 35.4006 nan 0.3896 -0.0325
## 3300 35.3258 nan 0.3896 -0.0454
## 3320 35.3097 nan 0.3896 -0.0410
## 3340 35.2573 nan 0.3896 -0.0743
## 3360 35.1747 nan 0.3896 -0.1555
## 3380 35.1081 nan 0.3896 -0.0566
## 3400 35.1279 nan 0.3896 -0.0905
## 3420 35.0406 nan 0.3896 -0.0422
## 3440 34.9268 nan 0.3896 -0.0547
## 3460 34.8506 nan 0.3896 -0.0650
## 3480 34.8232 nan 0.3896 -0.1557
## 3500 34.8129 nan 0.3896 -0.4201
## 3520 34.6681 nan 0.3896 -0.0877
## 3540 34.7046 nan 0.3896 -0.4916
## 3560 34.6142 nan 0.3896 -0.0295
## 3580 34.5245 nan 0.3896 -0.0654
## 3600 34.4668 nan 0.3896 -0.0863
## 3620 34.3957 nan 0.3896 -0.1030
## 3640 34.3557 nan 0.3896 -0.2238
## 3660 34.2600 nan 0.3896 -0.0848
## 3680 34.1688 nan 0.3896 -0.0413
## 3700 34.1401 nan 0.3896 -0.1209
## 3720 34.0873 nan 0.3896 -0.0301
## 3740 34.0117 nan 0.3896 -0.0984
## 3760 33.9812 nan 0.3896 0.0022
## 3780 33.8756 nan 0.3896 -0.0920
## 3800 33.7887 nan 0.3896 -0.0653
## 3820 33.7313 nan 0.3896 -0.0637
## 3840 33.6819 nan 0.3896 -0.1124
## 3860 33.6066 nan 0.3896 -0.0679
## 3880 33.5754 nan 0.3896 -0.0418
## 3900 33.5195 nan 0.3896 -0.0609
## 3920 33.4548 nan 0.3896 -0.1377
## 3940 33.4382 nan 0.3896 -0.0957
## 3960 33.3645 nan 0.3896 -0.1016
## 3980 33.2662 nan 0.3896 -0.0530
## 4000 33.2574 nan 0.3896 -0.0898
## 4020 33.1662 nan 0.3896 -0.0525
## 4040 33.1347 nan 0.3896 -0.0675
## 4060 33.0526 nan 0.3896 -0.0353
## 4080 32.9741 nan 0.3896 -0.0971
##
## Iter TrainDeviance ValidDeviance StepSize Improve
## 1 551.9731 nan 0.5470 638.5965
## 2 381.1555 nan 0.5470 164.1417
## 3 315.8714 nan 0.5470 61.7841
## 4 287.3096 nan 0.5470 28.0545
## 5 275.1857 nan 0.5470 11.5459
## 6 260.4648 nan 0.5470 13.0979
## 7 251.7576 nan 0.5470 7.2453
## 8 245.4612 nan 0.5470 5.5279
## 9 239.0924 nan 0.5470 4.9987
## 10 232.5708 nan 0.5470 5.4367
## 20 194.3271 nan 0.5470 3.1016
## 40 164.6481 nan 0.5470 0.8264
## 60 146.9274 nan 0.5470 -0.3204
## 80 133.5491 nan 0.5470 0.0071
## 100 125.9370 nan 0.5470 -0.0152
## 120 118.3852 nan 0.5470 -0.1809
## 140 113.2273 nan 0.5470 -0.4388
## 160 108.4859 nan 0.5470 -0.0915
## 180 103.6084 nan 0.5470 -0.3237
## 200 99.7271 nan 0.5470 -0.2485
## 220 96.5349 nan 0.5470 -0.2696
## 240 94.1205 nan 0.5470 -0.4432
## 260 91.3689 nan 0.5470 -0.1258
## 280 88.6295 nan 0.5470 -0.1268
## 300 86.6224 nan 0.5470 -0.2565
## 320 84.4482 nan 0.5470 -0.2164
## 340 82.4291 nan 0.5470 -0.1545
## 360 80.4927 nan 0.5470 -0.2878
## 380 78.6497 nan 0.5470 -0.1883
## 400 77.5114 nan 0.5470 -0.1680
## 420 76.2015 nan 0.5470 -0.2295
## 440 74.9412 nan 0.5470 -0.1248
## 460 73.7550 nan 0.5470 -0.4878
## 480 72.5179 nan 0.5470 -0.1957
## 500 71.7275 nan 0.5470 -0.4899
## 520 70.6845 nan 0.5470 -0.3163
## 540 69.6923 nan 0.5470 -0.1638
## 560 68.4473 nan 0.5470 -0.1358
## 580 67.7863 nan 0.5470 -0.2149
## 600 67.0501 nan 0.5470 -0.0414
## 620 66.1772 nan 0.5470 -0.4016
## 640 65.3527 nan 0.5470 -0.1379
## 660 64.5558 nan 0.5470 -0.1046
## 680 63.8072 nan 0.5470 0.0027
## 700 62.8732 nan 0.5470 -0.1552
## 720 62.3283 nan 0.5470 -0.2053
## 740 61.8211 nan 0.5470 -0.4524
## 760 61.1574 nan 0.5470 -0.1339
## 780 60.4313 nan 0.5470 -0.1831
## 800 59.8244 nan 0.5470 -0.1173
## 820 59.1227 nan 0.5470 -0.0616
## 840 58.4032 nan 0.5470 -0.0936
## 860 57.9002 nan 0.5470 -0.2349
## 880 57.0831 nan 0.5470 -0.0592
## 900 56.7251 nan 0.5470 -0.3941
## 920 56.0808 nan 0.5470 -0.1098
## 940 55.3472 nan 0.5470 -0.1270
## 960 55.1568 nan 0.5470 -0.0940
## 980 54.7041 nan 0.5470 -0.1750
## 1000 54.3076 nan 0.5470 -0.1086
## 1020 53.8875 nan 0.5470 -0.1628
## 1040 53.4368 nan 0.5470 -0.0800
## 1060 53.1067 nan 0.5470 -0.2687
## 1080 52.6952 nan 0.5470 -0.1260
## 1100 52.3135 nan 0.5470 -0.1394
## 1120 51.9353 nan 0.5470 -0.1390
## 1140 51.6637 nan 0.5470 0.0117
## 1160 51.4217 nan 0.5470 -0.1881
## 1180 51.0632 nan 0.5470 -0.0363
## 1200 50.7004 nan 0.5470 -0.1361
## 1220 50.3957 nan 0.5470 -0.1643
## 1240 49.9541 nan 0.5470 -0.2045
## 1260 49.7184 nan 0.5470 -0.0926
## 1280 49.4778 nan 0.5470 -0.2371
## 1300 49.2324 nan 0.5470 -0.2377
## 1320 48.8587 nan 0.5470 -0.2615
## 1340 48.5517 nan 0.5470 -0.2015
## 1360 48.2153 nan 0.5470 -0.1236
## 1380 47.9920 nan 0.5470 -0.1454
## 1400 47.6195 nan 0.5470 -0.0642
## 1420 47.5009 nan 0.5470 -0.0489
## 1440 47.4044 nan 0.5470 -0.0808
## 1460 47.0660 nan 0.5470 -0.2272
## 1480 46.9046 nan 0.5470 -0.1763
## 1500 46.5563 nan 0.5470 -0.1181
## 1520 46.2395 nan 0.5470 -0.1260
## 1540 45.9403 nan 0.5470 -0.2020
## 1560 45.7985 nan 0.5470 -0.3567
## 1580 45.4723 nan 0.5470 -0.1314
## 1600 45.3330 nan 0.5470 -0.2233
## 1620 45.0521 nan 0.5470 -0.1624
## 1640 44.8346 nan 0.5470 -0.1625
## 1660 44.4841 nan 0.5470 -0.1541
## 1680 44.3107 nan 0.5470 -0.2384
## 1700 44.1323 nan 0.5470 -0.1604
## 1720 43.9816 nan 0.5470 -0.1296
## 1740 43.8142 nan 0.5470 -0.0477
## 1760 43.8027 nan 0.5470 -0.6810
## 1780 43.5847 nan 0.5470 -0.1131
## 1800 43.5105 nan 0.5470 -0.1025
## 1820 43.4377 nan 0.5470 -0.3473
## 1840 43.2952 nan 0.5470 -0.2029
## 1860 42.8899 nan 0.5470 -0.0207
## 1880 42.6106 nan 0.5470 -0.1103
## 1900 42.3187 nan 0.5470 -0.2466
## 1920 42.0889 nan 0.5470 -0.0715
## 1940 41.8960 nan 0.5470 -0.2247
## 1960 41.7147 nan 0.5470 -0.1536
## 1980 41.5955 nan 0.5470 -0.1095
## 2000 41.4141 nan 0.5470 -0.1985
## 2020 41.2675 nan 0.5470 -0.1558
## 2040 41.1475 nan 0.5470 -0.0928
## 2060 40.9951 nan 0.5470 -0.1342
## 2080 40.8058 nan 0.5470 -0.1443
## 2100 40.6872 nan 0.5470 -0.1671
## 2120 40.5094 nan 0.5470 -0.0588
## 2140 40.4106 nan 0.5470 -0.0386
## 2160 40.3516 nan 0.5470 -0.2805
## 2180 40.2182 nan 0.5470 -0.1903
## 2200 40.1394 nan 0.5470 -0.2981
## 2220 39.9887 nan 0.5470 -0.1629
## 2240 39.8891 nan 0.5470 -0.1952
## 2260 39.7712 nan 0.5470 -0.2200
## 2280 39.6615 nan 0.5470 -0.1521
## 2300 39.5066 nan 0.5470 -0.0853
## 2320 39.4965 nan 0.5470 -0.1469
## 2340 39.3466 nan 0.5470 -0.0986
## 2360 39.2249 nan 0.5470 -0.0744
## 2380 39.2696 nan 0.5470 -0.2722
## 2400 39.0490 nan 0.5470 -0.1614
## 2420 38.9093 nan 0.5470 -0.1274
## 2440 38.9146 nan 0.5470 -0.1227
## 2460 38.8017 nan 0.5470 -0.1196
## 2480 38.6315 nan 0.5470 -0.1009
## 2500 38.5830 nan 0.5470 -0.2261
## 2520 38.4996 nan 0.5470 -0.2246
## 2540 38.3495 nan 0.5470 -0.0943
## 2560 38.2436 nan 0.5470 -0.3023
## 2580 38.2413 nan 0.5470 -0.2834
## 2600 38.1129 nan 0.5470 -0.3221
## 2620 37.9623 nan 0.5470 -0.1470
## 2640 37.8007 nan 0.5470 -0.2301
## 2660 37.8060 nan 0.5470 -0.3747
## 2680 37.7995 nan 0.5470 -0.1252
## 2700 37.7880 nan 0.5470 -0.5029
## 2720 37.6653 nan 0.5470 -0.1357
## 2740 37.5853 nan 0.5470 -0.1693
## 2760 37.3747 nan 0.5470 -0.1392
## 2780 37.2957 nan 0.5470 -0.1396
## 2800 37.3461 nan 0.5470 -0.2411
## 2820 37.2597 nan 0.5470 -0.1097
## 2840 37.1985 nan 0.5470 -0.2134
## 2860 37.0586 nan 0.5470 -0.0822
## 2880 36.9963 nan 0.5470 -0.2091
## 2900 36.9203 nan 0.5470 -0.1675
## 2920 36.8555 nan 0.5470 -0.1453
## 2940 36.7506 nan 0.5470 -0.0930
## 2960 36.6430 nan 0.5470 -0.0885
## 2980 36.7613 nan 0.5470 -0.0919
## 3000 36.6627 nan 0.5470 -0.1375
## 3020 36.5344 nan 0.5470 -0.1845
## 3040 36.5405 nan 0.5470 -0.1252
## 3060 36.5445 nan 0.5470 -0.2860
## 3080 36.4132 nan 0.5470 -0.1219
## 3100 36.2556 nan 0.5470 -0.0779
## 3120 36.3051 nan 0.5470 -0.1318
## 3140 36.2897 nan 0.5470 -0.1977
## 3160 36.2187 nan 0.5470 -0.0748
## 3180 36.0807 nan 0.5470 -0.1682
## 3200 35.9852 nan 0.5470 -0.0433
## 3220 36.0114 nan 0.5470 -0.2199
## 3240 35.8993 nan 0.5470 -0.1885
## 3260 35.8437 nan 0.5470 -0.1442
## 3280 35.7896 nan 0.5470 -0.0613
## 3300 35.7146 nan 0.5470 -0.1829
## 3320 35.5908 nan 0.5470 -0.1382
## 3340 35.5610 nan 0.5470 -0.1563
## 3360 35.5262 nan 0.5470 -0.2959
## 3380 35.3544 nan 0.5470 -0.0131
## 3400 35.3355 nan 0.5470 -0.1116
## 3420 35.2643 nan 0.5470 -0.1664
## 3440 35.1757 nan 0.5470 -0.1161
## 3460 35.0879 nan 0.5470 -0.1496
## 3480 35.0917 nan 0.5470 -0.1281
## 3489 35.0733 nan 0.5470 -0.1099
##
## Iter TrainDeviance ValidDeviance StepSize Improve
## 1 824.5032 nan 0.2306 358.7594
## 2 602.8762 nan 0.2306 218.3947
## 3 464.0081 nan 0.2306 139.9403
## 4 373.1170 nan 0.2306 87.2792
## 5 314.6836 nan 0.2306 56.7475
## 6 275.4414 nan 0.2306 37.9960
## 7 248.6566 nan 0.2306 25.9397
## 8 230.2933 nan 0.2306 17.7199
## 9 216.9252 nan 0.2306 11.9727
## 10 206.7317 nan 0.2306 9.3349
## 20 163.8796 nan 0.2306 1.7141
## 40 133.0524 nan 0.2306 0.7639
## 60 116.7856 nan 0.2306 0.3315
## 80 106.4404 nan 0.2306 0.0371
## 100 97.3825 nan 0.2306 0.0168
## 120 90.3686 nan 0.2306 -0.1658
## 140 84.9621 nan 0.2306 -0.0848
## 160 79.9917 nan 0.2306 -0.0244
## 180 76.1899 nan 0.2306 -0.1521
## 200 72.1666 nan 0.2306 -0.1482
## 220 69.0918 nan 0.2306 -0.1621
## 240 66.3786 nan 0.2306 -0.2010
## 260 63.6474 nan 0.2306 -0.1207
## 280 61.1302 nan 0.2306 -0.1167
## 300 58.8473 nan 0.2306 -0.0896
## 320 57.0501 nan 0.2306 -0.2334
## 340 55.3198 nan 0.2306 -0.0880
## 360 53.7088 nan 0.2306 -0.1095
## 380 52.1479 nan 0.2306 -0.3058
## 400 50.3986 nan 0.2306 -0.2488
## 420 49.1037 nan 0.2306 -0.1572
## 440 48.0553 nan 0.2306 -0.1990
## 460 46.9009 nan 0.2306 -0.2479
## 480 45.9653 nan 0.2306 -0.1983
## 500 44.9914 nan 0.2306 -0.1008
## 520 44.1757 nan 0.2306 -0.1013
## 540 43.2599 nan 0.2306 -0.1060
## 560 42.4597 nan 0.2306 -0.2213
## 580 41.7980 nan 0.2306 -0.2266
## 600 41.1184 nan 0.2306 -0.3225
## 620 40.3244 nan 0.2306 -0.2067
## 640 39.7375 nan 0.2306 -0.1362
## 660 39.2380 nan 0.2306 -0.1679
## 680 38.5639 nan 0.2306 -0.1238
## 700 38.1501 nan 0.2306 -0.1340
## 720 37.5051 nan 0.2306 -0.0775
## 740 37.0586 nan 0.2306 -0.1470
## 760 36.7057 nan 0.2306 -0.1541
## 780 36.2024 nan 0.2306 -0.1019
## 800 35.8285 nan 0.2306 -0.0979
## 820 35.4177 nan 0.2306 -0.1742
## 840 34.9678 nan 0.2306 -0.0896
## 860 34.5419 nan 0.2306 -0.2280
## 880 34.1168 nan 0.2306 -0.1592
## 900 33.7109 nan 0.2306 -0.1570
## 920 33.3359 nan 0.2306 -0.0764
## 940 32.9939 nan 0.2306 -0.1531
## 960 32.7770 nan 0.2306 -0.0818
## 980 32.3704 nan 0.2306 -0.0311
## 1000 32.1364 nan 0.2306 -0.1393
## 1020 31.8902 nan 0.2306 -0.1330
## 1040 31.6040 nan 0.2306 -0.1325
## 1060 31.4326 nan 0.2306 -0.2165
## 1080 31.2036 nan 0.2306 -0.1640
## 1100 30.9305 nan 0.2306 -0.1841
## 1120 30.7497 nan 0.2306 -0.1274
## 1140 30.5859 nan 0.2306 -0.1041
## 1160 30.4050 nan 0.2306 -0.1063
## 1180 30.2025 nan 0.2306 -0.1692
## 1200 30.0246 nan 0.2306 -0.1688
## 1220 29.8867 nan 0.2306 -0.1931
## 1240 29.6317 nan 0.2306 -0.0471
## 1260 29.4578 nan 0.2306 -0.1213
## 1280 29.3360 nan 0.2306 -0.1516
## 1300 29.2298 nan 0.2306 -0.0640
## 1320 29.0652 nan 0.2306 -0.1261
## 1340 28.9356 nan 0.2306 -0.2399
## 1360 28.7919 nan 0.2306 -0.1077
## 1380 28.6566 nan 0.2306 -0.1116
## 1400 28.5466 nan 0.2306 -0.1442
## 1420 28.3952 nan 0.2306 -0.1197
## 1440 28.3147 nan 0.2306 -0.2139
## 1460 28.1220 nan 0.2306 -0.1787
## 1480 27.9224 nan 0.2306 -0.1238
## 1500 27.7999 nan 0.2306 -0.1230
## 1520 27.6762 nan 0.2306 -0.1884
## 1540 27.5669 nan 0.2306 -0.1186
## 1560 27.4850 nan 0.2306 -0.1428
## 1580 27.3843 nan 0.2306 -0.0660
## 1600 27.3056 nan 0.2306 -0.0864
## 1620 27.1786 nan 0.2306 -0.1350
## 1640 27.0494 nan 0.2306 -0.0375
## 1660 27.0183 nan 0.2306 -0.1578
## 1680 26.9686 nan 0.2306 -0.1544
## 1700 26.8209 nan 0.2306 -0.2855
## 1720 26.6955 nan 0.2306 -0.1143
## 1740 26.5769 nan 0.2306 -0.1382
## 1760 26.4889 nan 0.2306 -0.1159
## 1780 26.4049 nan 0.2306 -0.0947
## 1800 26.3146 nan 0.2306 -0.0891
## 1820 26.2605 nan 0.2306 -0.0572
## 1840 26.2008 nan 0.2306 -0.1500
## 1860 26.1285 nan 0.2306 -0.1846
## 1880 26.1385 nan 0.2306 -0.2176
## 1900 25.9891 nan 0.2306 -0.1341
## 1920 25.9143 nan 0.2306 -0.1143
## 1940 25.8488 nan 0.2306 -0.0886
## 1960 25.7927 nan 0.2306 -0.1476
## 1980 25.7497 nan 0.2306 -0.1884
## 2000 25.6909 nan 0.2306 -0.1705
## 2020 25.5852 nan 0.2306 -0.0803
## 2040 25.4981 nan 0.2306 -0.1208
## 2060 25.4482 nan 0.2306 -0.1295
## 2080 25.3058 nan 0.2306 -0.1088
## 2100 25.2177 nan 0.2306 -0.0869
## 2120 25.1747 nan 0.2306 -0.0986
## 2140 25.1074 nan 0.2306 -0.0651
## 2160 25.0492 nan 0.2306 -0.0526
## 2180 24.9790 nan 0.2306 -0.0638
## 2200 24.9555 nan 0.2306 -0.0537
## 2220 24.8798 nan 0.2306 -0.0514
## 2240 24.7932 nan 0.2306 -0.0948
## 2260 24.7678 nan 0.2306 -0.1029
## 2280 24.7128 nan 0.2306 -0.0798
## 2300 24.6902 nan 0.2306 -0.1464
## 2320 24.6225 nan 0.2306 -0.0694
## 2340 24.6463 nan 0.2306 -0.0760
## 2360 24.5509 nan 0.2306 -0.2214
## 2380 24.4958 nan 0.2306 -0.1149
## 2400 24.4500 nan 0.2306 -0.0977
## 2420 24.4441 nan 0.2306 -0.1574
## 2440 24.3766 nan 0.2306 -0.1606
## 2460 24.2865 nan 0.2306 -0.0496
## 2480 24.2534 nan 0.2306 -0.0645
## 2500 24.1952 nan 0.2306 -0.1443
## 2520 24.1762 nan 0.2306 -0.1407
## 2540 24.1635 nan 0.2306 -0.1334
## 2560 24.1188 nan 0.2306 -0.1231
## 2580 24.1100 nan 0.2306 -0.0878
## 2600 24.0546 nan 0.2306 -0.1168
## 2620 24.0522 nan 0.2306 -0.1046
## 2640 23.9633 nan 0.2306 -0.2214
## 2660 23.9170 nan 0.2306 -0.1343
## 2680 23.8495 nan 0.2306 -0.0995
## 2700 23.7887 nan 0.2306 -0.0992
## 2720 23.7930 nan 0.2306 -0.1271
## 2740 23.7055 nan 0.2306 -0.1114
## 2760 23.7048 nan 0.2306 -0.1145
## 2780 23.6803 nan 0.2306 -0.2032
## 2800 23.6571 nan 0.2306 -0.1405
## 2820 23.6065 nan 0.2306 -0.2271
## 2840 23.5749 nan 0.2306 -0.0958
## 2860 23.5475 nan 0.2306 -0.1426
## 2880 23.5379 nan 0.2306 -0.1520
## 2900 23.4992 nan 0.2306 -0.1104
## 2920 23.4745 nan 0.2306 -0.0425
## 2940 23.4463 nan 0.2306 -0.1049
## 2960 23.4108 nan 0.2306 -0.0249
## 2980 23.3516 nan 0.2306 -0.0429
## 3000 23.3262 nan 0.2306 -0.0776
## 3020 23.3207 nan 0.2306 -0.1422
## 3040 23.2846 nan 0.2306 -0.1803
## 3060 23.2596 nan 0.2306 -0.1003
## 3080 23.2887 nan 0.2306 -0.1694
## 3100 23.2525 nan 0.2306 -0.1736
## 3120 23.1977 nan 0.2306 -0.1343
## 3140 23.2284 nan 0.2306 -0.0869
## 3160 23.1945 nan 0.2306 -0.2439
## 3180 23.1641 nan 0.2306 -0.2110
## 3200 23.0972 nan 0.2306 -0.1107
## 3220 23.0840 nan 0.2306 -0.1662
## 3240 23.0224 nan 0.2306 -0.1066
## 3260 23.0106 nan 0.2306 -0.2042
## 3280 23.0024 nan 0.2306 -0.1358
## 3300 22.9757 nan 0.2306 -0.1037
## 3320 22.9466 nan 0.2306 -0.0573
## 3340 22.8861 nan 0.2306 -0.0984
## 3360 22.8520 nan 0.2306 -0.1346
## 3380 22.8440 nan 0.2306 -0.1262
## 3400 22.8351 nan 0.2306 -0.1342
## 3420 22.7839 nan 0.2306 -0.1059
## 3440 22.8117 nan 0.2306 -0.1889
## 3460 22.7635 nan 0.2306 -0.1571
## 3480 22.7803 nan 0.2306 -0.1192
## 3500 22.7506 nan 0.2306 -0.0922
## 3520 22.7167 nan 0.2306 -0.1205
## 3540 22.7646 nan 0.2306 -0.3213
## 3560 22.7455 nan 0.2306 -0.1089
## 3580 22.6815 nan 0.2306 -0.0948
## 3600 22.6599 nan 0.2306 -0.1139
## 3620 22.6339 nan 0.2306 -0.1306
## 3640 22.5907 nan 0.2306 -0.0924
## 3660 22.5897 nan 0.2306 -0.0755
## 3680 22.5878 nan 0.2306 -0.0389
## 3700 22.5453 nan 0.2306 -0.0883
## 3720 22.5001 nan 0.2306 -0.0678
##
## Iter TrainDeviance ValidDeviance StepSize Improve
## 1 676.7874 nan 0.3896 503.0298
## 2 462.1272 nan 0.3896 217.5228
## 3 369.3819 nan 0.3896 91.6676
## 4 321.1326 nan 0.3896 48.9746
## 5 292.0548 nan 0.3896 29.0610
## 6 276.6196 nan 0.3896 15.4489
## 7 267.1979 nan 0.3896 8.8064
## 8 256.3657 nan 0.3896 10.2001
## 9 249.4938 nan 0.3896 6.1156
## 10 241.7859 nan 0.3896 7.3177
## 20 204.6758 nan 0.3896 1.9954
## 40 171.9567 nan 0.3896 0.6936
## 60 154.4838 nan 0.3896 0.1518
## 80 141.4603 nan 0.3896 0.1641
## 100 131.6954 nan 0.3896 -0.1131
## 120 124.0250 nan 0.3896 -0.0188
## 140 118.8261 nan 0.3896 -0.0867
## 160 113.3610 nan 0.3896 -0.0144
## 180 109.7572 nan 0.3896 -0.1600
## 200 105.9059 nan 0.3896 -0.3059
## 220 101.8396 nan 0.3896 -0.1546
## 240 97.9844 nan 0.3896 0.0017
## 260 95.0951 nan 0.3896 -0.0098
## 280 92.9656 nan 0.3896 -0.1020
## 300 91.5154 nan 0.3896 -0.1820
## 320 89.6842 nan 0.3896 -0.0738
## 340 87.4336 nan 0.3896 -0.1016
## 360 85.6375 nan 0.3896 -0.0074
## 380 83.5569 nan 0.3896 -0.0381
## 400 81.5011 nan 0.3896 -0.1940
## 420 80.2534 nan 0.3896 -0.2900
## 440 78.9520 nan 0.3896 -0.3901
## 460 77.2449 nan 0.3896 -0.1287
## 480 75.7496 nan 0.3896 -0.1397
## 500 74.6485 nan 0.3896 -0.0332
## 520 73.4894 nan 0.3896 -0.1106
## 540 72.4329 nan 0.3896 -0.2478
## 560 71.5887 nan 0.3896 -0.1621
## 580 70.8836 nan 0.3896 -0.2270
## 600 69.9890 nan 0.3896 -0.1757
## 620 69.0725 nan 0.3896 -0.0638
## 640 68.0818 nan 0.3896 -0.0799
## 660 67.1039 nan 0.3896 -0.0329
## 680 66.4275 nan 0.3896 -0.2072
## 700 65.6793 nan 0.3896 -0.2348
## 720 64.7027 nan 0.3896 -0.1839
## 740 64.1165 nan 0.3896 -0.0104
## 760 63.4867 nan 0.3896 -0.0646
## 780 62.8271 nan 0.3896 -0.0672
## 800 62.1435 nan 0.3896 -0.1374
## 820 61.5516 nan 0.3896 -0.1716
## 840 61.1191 nan 0.3896 -0.2039
## 860 60.5658 nan 0.3896 -0.1957
## 880 60.1272 nan 0.3896 -0.0647
## 900 59.6326 nan 0.3896 -0.0670
## 920 59.2187 nan 0.3896 -0.0671
## 940 58.7571 nan 0.3896 -0.0759
## 960 58.3047 nan 0.3896 -0.0923
## 980 57.7488 nan 0.3896 -0.1091
## 1000 57.1954 nan 0.3896 -0.0849
## 1020 56.6139 nan 0.3896 -0.0456
## 1040 56.1948 nan 0.3896 -0.0731
## 1060 55.7054 nan 0.3896 -0.1210
## 1080 55.1929 nan 0.3896 -0.1745
## 1100 54.7042 nan 0.3896 -0.0754
## 1120 54.3278 nan 0.3896 -0.1089
## 1140 53.9743 nan 0.3896 -0.1038
## 1160 53.5408 nan 0.3896 -0.1734
## 1180 53.1545 nan 0.3896 -0.1132
## 1200 52.7538 nan 0.3896 -0.1956
## 1220 52.4848 nan 0.3896 -0.1043
## 1240 52.1655 nan 0.3896 -0.1915
## 1260 51.8808 nan 0.3896 -0.0835
## 1280 51.4990 nan 0.3896 -0.0639
## 1300 51.0864 nan 0.3896 -0.1306
## 1320 50.8608 nan 0.3896 -0.0614
## 1340 50.4388 nan 0.3896 -0.0872
## 1360 50.0544 nan 0.3896 -0.1688
## 1380 49.8039 nan 0.3896 -0.0831
## 1400 49.3310 nan 0.3896 -0.0684
## 1420 49.0754 nan 0.3896 -0.0810
## 1440 48.5394 nan 0.3896 -0.1353
## 1460 48.2373 nan 0.3896 -0.0448
## 1480 47.8167 nan 0.3896 -0.0651
## 1500 47.4877 nan 0.3896 -0.1033
## 1520 47.1899 nan 0.3896 -0.1368
## 1540 46.9605 nan 0.3896 -0.1168
## 1560 46.7671 nan 0.3896 -0.1675
## 1580 46.5418 nan 0.3896 -0.1274
## 1600 46.2932 nan 0.3896 -0.0774
## 1620 45.9785 nan 0.3896 -0.0975
## 1640 45.7235 nan 0.3896 -0.1089
## 1660 45.4349 nan 0.3896 -0.0717
## 1680 45.2012 nan 0.3896 -0.1227
## 1700 45.0619 nan 0.3896 -0.0673
## 1720 44.8754 nan 0.3896 -0.0701
## 1740 44.5822 nan 0.3896 -0.0016
## 1760 44.4372 nan 0.3896 -0.2114
## 1780 44.2204 nan 0.3896 -0.0916
## 1800 44.0977 nan 0.3896 -0.1041
## 1820 43.8926 nan 0.3896 -0.0844
## 1840 43.7616 nan 0.3896 -0.1153
## 1860 43.4829 nan 0.3896 -0.1127
## 1880 43.3235 nan 0.3896 -0.0672
## 1900 43.0746 nan 0.3896 -0.0592
## 1920 42.8841 nan 0.3896 -0.0877
## 1940 42.6049 nan 0.3896 -0.1000
## 1960 42.3771 nan 0.3896 -0.0960
## 1980 42.2022 nan 0.3896 -0.1194
## 2000 42.0165 nan 0.3896 -0.0727
## 2020 41.9102 nan 0.3896 -0.0669
## 2040 41.7578 nan 0.3896 -0.0751
## 2060 41.6062 nan 0.3896 -0.1005
## 2080 41.4870 nan 0.3896 -0.0751
## 2100 41.3008 nan 0.3896 -0.1438
## 2120 41.2152 nan 0.3896 -0.0725
## 2140 41.0719 nan 0.3896 -0.0947
## 2160 40.9667 nan 0.3896 -0.1601
## 2180 40.8411 nan 0.3896 -0.0374
## 2200 40.7262 nan 0.3896 -0.1341
## 2220 40.6091 nan 0.3896 -0.2143
## 2240 40.4798 nan 0.3896 -0.0928
## 2260 40.2775 nan 0.3896 -0.1040
## 2280 40.1822 nan 0.3896 -0.1540
## 2300 40.0495 nan 0.3896 -0.0409
## 2320 39.8311 nan 0.3896 -0.2209
## 2340 39.7834 nan 0.3896 -0.1052
## 2360 39.6512 nan 0.3896 -0.1208
## 2380 39.4127 nan 0.3896 -0.0611
## 2400 39.3839 nan 0.3896 -0.1135
## 2420 39.2398 nan 0.3896 -0.0600
## 2440 39.0575 nan 0.3896 -0.0679
## 2460 38.9582 nan 0.3896 -0.0562
## 2480 38.8427 nan 0.3896 -0.0873
## 2500 38.8203 nan 0.3896 -0.1018
## 2520 38.6111 nan 0.3896 -0.0779
## 2540 38.5028 nan 0.3896 -0.1741
## 2560 38.3847 nan 0.3896 -0.0558
## 2580 38.2446 nan 0.3896 -0.0542
## 2600 38.1379 nan 0.3896 -0.1146
## 2620 37.9773 nan 0.3896 -0.1078
## 2640 37.8399 nan 0.3896 -0.0330
## 2660 37.6682 nan 0.3896 -0.0808
## 2680 37.5244 nan 0.3896 -0.1111
## 2700 37.4741 nan 0.3896 -0.0376
## 2720 37.3291 nan 0.3896 -0.0754
## 2740 37.1783 nan 0.3896 -0.1129
## 2760 37.0910 nan 0.3896 -0.0882
## 2780 37.0480 nan 0.3896 -0.0953
## 2800 36.8962 nan 0.3896 -0.1007
## 2820 36.7865 nan 0.3896 -0.0665
## 2840 36.6619 nan 0.3896 -0.0681
## 2860 36.6333 nan 0.3896 -0.0634
## 2880 36.5219 nan 0.3896 -0.1025
## 2900 36.4865 nan 0.3896 -0.0564
## 2920 36.4256 nan 0.3896 -0.1745
## 2940 36.3368 nan 0.3896 -0.1579
## 2960 36.2459 nan 0.3896 -0.1270
## 2980 36.1531 nan 0.3896 -0.1513
## 3000 36.0341 nan 0.3896 -0.1159
## 3020 35.9299 nan 0.3896 -0.0661
## 3040 35.8598 nan 0.3896 -0.0791
## 3060 35.7775 nan 0.3896 -0.1142
## 3080 35.6993 nan 0.3896 -0.0720
## 3100 35.6068 nan 0.3896 -0.1135
## 3120 35.5469 nan 0.3896 -0.1477
## 3140 35.4928 nan 0.3896 -0.0746
## 3160 35.4537 nan 0.3896 -0.1369
## 3180 35.3915 nan 0.3896 -0.0891
## 3200 35.2540 nan 0.3896 -0.1028
## 3220 35.1759 nan 0.3896 -0.1115
## 3240 35.0391 nan 0.3896 -0.0677
## 3260 34.9902 nan 0.3896 -0.1324
## 3280 34.9695 nan 0.3896 -0.0647
## 3300 34.8306 nan 0.3896 -0.0895
## 3320 34.7354 nan 0.3896 -0.0433
## 3340 34.7117 nan 0.3896 -0.2067
## 3360 34.6097 nan 0.3896 -0.0957
## 3380 34.5283 nan 0.3896 -0.1152
## 3400 34.4839 nan 0.3896 -0.1711
## 3420 34.3977 nan 0.3896 -0.0642
## 3440 34.3447 nan 0.3896 -0.1031
## 3460 34.3113 nan 0.3896 -0.1636
## 3480 34.1802 nan 0.3896 -0.0445
## 3500 34.1486 nan 0.3896 -0.0365
## 3520 34.0548 nan 0.3896 -0.0754
## 3540 33.9576 nan 0.3896 -0.1056
## 3560 33.9306 nan 0.3896 -0.0829
## 3580 33.9437 nan 0.3896 -0.3327
## 3600 33.7716 nan 0.3896 -0.1081
## 3620 33.6881 nan 0.3896 -0.0497
## 3640 33.6302 nan 0.3896 -0.0662
## 3660 33.6233 nan 0.3896 -0.1971
## 3680 33.5466 nan 0.3896 -0.0810
## 3700 33.4634 nan 0.3896 -0.0638
## 3720 33.4878 nan 0.3896 -0.1659
## 3740 33.4348 nan 0.3896 -0.0971
## 3760 33.3554 nan 0.3896 -0.0899
## 3780 33.2806 nan 0.3896 -0.0373
## 3800 33.2320 nan 0.3896 -0.0573
## 3820 33.2227 nan 0.3896 -0.0436
## 3840 33.1606 nan 0.3896 -0.1817
## 3860 33.0350 nan 0.3896 -0.0389
## 3880 32.9366 nan 0.3896 -0.1747
## 3900 32.8574 nan 0.3896 -0.0962
## 3920 32.8301 nan 0.3896 -0.0751
## 3940 32.7757 nan 0.3896 -0.0657
## 3960 32.6493 nan 0.3896 -0.0313
## 3980 32.6333 nan 0.3896 -0.0686
## 4000 32.5437 nan 0.3896 -0.1273
## 4020 32.5393 nan 0.3896 -0.2766
## 4040 32.4834 nan 0.3896 -0.1251
## 4060 32.4715 nan 0.3896 -0.0828
## 4080 32.3601 nan 0.3896 -0.0932
##
## Iter TrainDeviance ValidDeviance StepSize Improve
## 1 546.5661 nan 0.5470 641.0665
## 2 376.3570 nan 0.5470 170.3941
## 3 311.1084 nan 0.5470 64.0566
## 4 283.3075 nan 0.5470 22.7322
## 5 269.2665 nan 0.5470 12.1286
## 6 258.0019 nan 0.5470 9.3113
## 7 243.7206 nan 0.5470 13.7227
## 8 236.8729 nan 0.5470 5.9110
## 9 231.7542 nan 0.5470 4.1337
## 10 226.1930 nan 0.5470 5.1666
## 20 194.3862 nan 0.5470 1.9624
## 40 163.5186 nan 0.5470 0.9037
## 60 145.6874 nan 0.5470 -0.4706
## 80 136.6357 nan 0.5470 -0.8937
## 100 128.4355 nan 0.5470 -0.5505
## 120 122.2153 nan 0.5470 -0.2268
## 140 115.7605 nan 0.5470 -0.4969
## 160 111.0838 nan 0.5470 -0.0857
## 180 107.2105 nan 0.5470 -0.2006
## 200 102.9258 nan 0.5470 -0.2066
## 220 100.1367 nan 0.5470 -0.3513
## 240 97.2569 nan 0.5470 -0.2090
## 260 94.2185 nan 0.5470 -0.3163
## 280 92.2319 nan 0.5470 -0.2763
## 300 90.0157 nan 0.5470 -0.0852
## 320 87.7563 nan 0.5470 -0.2412
## 340 85.3624 nan 0.5470 0.0670
## 360 83.4321 nan 0.5470 -0.4163
## 380 81.7713 nan 0.5470 -0.3095
## 400 80.0384 nan 0.5470 -0.1843
## 420 78.4607 nan 0.5470 -0.2436
## 440 76.8827 nan 0.5470 -0.2332
## 460 75.5269 nan 0.5470 -0.1339
## 480 73.9243 nan 0.5470 -0.1011
## 500 72.5162 nan 0.5470 -0.1849
## 520 71.3064 nan 0.5470 -0.1424
## 540 70.2495 nan 0.5470 -0.2111
## 560 69.2767 nan 0.5470 -0.1808
## 580 68.2879 nan 0.5470 -0.3138
## 600 67.5192 nan 0.5470 -0.4171
## 620 66.5601 nan 0.5470 -0.2388
## 640 65.6637 nan 0.5470 -0.0992
## 660 64.7229 nan 0.5470 -0.2879
## 680 64.0665 nan 0.5470 -0.2142
## 700 63.3618 nan 0.5470 -0.0640
## 720 62.4720 nan 0.5470 -0.2646
## 740 61.6768 nan 0.5470 -0.1569
## 760 60.8921 nan 0.5470 -0.1717
## 780 60.1663 nan 0.5470 -0.2218
## 800 59.5106 nan 0.5470 -0.2020
## 820 58.8745 nan 0.5470 -0.0819
## 840 58.2365 nan 0.5470 -0.2009
## 860 57.7834 nan 0.5470 -0.5404
## 880 57.2998 nan 0.5470 -0.1650
## 900 56.7681 nan 0.5470 -0.0522
## 920 56.2528 nan 0.5470 -0.1455
## 940 56.0001 nan 0.5470 -0.2303
## 960 55.6396 nan 0.5470 -0.1041
## 980 55.2729 nan 0.5470 -0.1234
## 1000 54.9231 nan 0.5470 -0.2537
## 1020 54.5413 nan 0.5470 -0.2304
## 1040 54.0163 nan 0.5470 -0.1779
## 1060 53.6049 nan 0.5470 -0.1047
## 1080 53.1617 nan 0.5470 -0.0618
## 1100 52.8327 nan 0.5470 -0.2596
## 1120 52.4054 nan 0.5470 -0.0512
## 1140 52.0399 nan 0.5470 -0.1604
## 1160 51.7141 nan 0.5470 -0.2805
## 1180 51.3139 nan 0.5470 -0.2143
## 1200 50.9560 nan 0.5470 -0.2674
## 1220 50.7072 nan 0.5470 -0.2872
## 1240 50.2833 nan 0.5470 -0.1485
## 1260 49.9520 nan 0.5470 -0.2206
## 1280 49.5444 nan 0.5470 -0.1601
## 1300 49.2284 nan 0.5470 -0.0329
## 1320 48.8895 nan 0.5470 -0.1568
## 1340 48.5203 nan 0.5470 -0.1167
## 1360 48.2126 nan 0.5470 -0.1167
## 1380 47.9532 nan 0.5470 -0.1552
## 1400 47.6969 nan 0.5470 -0.1928
## 1420 47.4716 nan 0.5470 -0.2516
## 1440 47.2213 nan 0.5470 -0.2152
## 1460 47.1133 nan 0.5470 -0.2246
## 1480 46.7965 nan 0.5470 -0.0723
## 1500 46.6405 nan 0.5470 -0.2787
## 1520 46.2043 nan 0.5470 -0.1304
## 1540 46.0356 nan 0.5470 -0.3637
## 1560 45.8195 nan 0.5470 -0.0614
## 1580 45.7218 nan 0.5470 -0.6509
## 1600 45.4390 nan 0.5470 -0.1670
## 1620 45.2994 nan 0.5470 -0.2902
## 1640 44.9795 nan 0.5470 -0.1743
## 1660 44.8234 nan 0.5470 -0.2952
## 1680 44.4551 nan 0.5470 -0.1528
## 1700 44.3912 nan 0.5470 -0.3037
## 1720 44.0968 nan 0.5470 -0.1346
## 1740 43.9784 nan 0.5470 -0.2377
## 1760 43.7942 nan 0.5470 -0.2863
## 1780 43.5378 nan 0.5470 -0.0925
## 1800 43.3838 nan 0.5470 -0.1641
## 1820 43.1275 nan 0.5470 -0.0948
## 1840 42.9637 nan 0.5470 -0.1626
## 1860 42.7432 nan 0.5470 -0.0886
## 1880 42.5478 nan 0.5470 -0.1150
## 1900 42.3355 nan 0.5470 -0.0525
## 1920 42.2099 nan 0.5470 -0.1942
## 1940 42.0787 nan 0.5470 -0.1890
## 1960 41.9156 nan 0.5470 -0.1586
## 1980 41.8686 nan 0.5470 -0.2626
## 2000 41.8021 nan 0.5470 -0.4455
## 2020 41.5710 nan 0.5470 -0.1698
## 2040 41.3957 nan 0.5470 -0.1744
## 2060 41.2927 nan 0.5470 -0.1867
## 2080 41.1829 nan 0.5470 -0.2127
## 2100 40.9783 nan 0.5470 -0.1668
## 2120 40.6914 nan 0.5470 -0.0516
## 2140 40.5860 nan 0.5470 -0.1782
## 2160 40.4169 nan 0.5470 -0.1933
## 2180 40.4164 nan 0.5470 -0.4715
## 2200 40.1695 nan 0.5470 -0.1923
## 2220 40.1263 nan 0.5470 -0.1078
## 2240 40.0211 nan 0.5470 -0.1873
## 2260 39.8246 nan 0.5470 -0.1064
## 2280 39.6909 nan 0.5470 -0.1938
## 2300 39.5407 nan 0.5470 -0.1404
## 2320 39.4408 nan 0.5470 -0.1769
## 2340 39.3522 nan 0.5470 -0.1558
## 2360 39.2391 nan 0.5470 -0.1730
## 2380 39.2133 nan 0.5470 -0.0453
## 2400 39.1104 nan 0.5470 -0.2066
## 2420 39.0067 nan 0.5470 -0.0802
## 2440 38.7941 nan 0.5470 -0.1459
## 2460 38.5926 nan 0.5470 -0.1749
## 2480 38.6276 nan 0.5470 -0.2508
## 2500 38.4237 nan 0.5470 -0.1558
## 2520 38.3170 nan 0.5470 -0.1633
## 2540 38.1402 nan 0.5470 -0.2759
## 2560 38.0954 nan 0.5470 -0.0568
## 2580 37.9986 nan 0.5470 -0.0939
## 2600 38.0530 nan 0.5470 -0.5297
## 2620 37.7259 nan 0.5470 -0.1556
## 2640 37.6480 nan 0.5470 -0.1346
## 2660 37.5463 nan 0.5470 -0.1037
## 2680 37.4748 nan 0.5470 -0.4596
## 2700 37.3463 nan 0.5470 -0.1313
## 2720 37.2085 nan 0.5470 -0.0976
## 2740 37.0329 nan 0.5470 -0.1684
## 2760 36.9484 nan 0.5470 -0.0875
## 2780 36.8370 nan 0.5470 -0.1809
## 2800 36.7521 nan 0.5470 -0.2574
## 2820 36.8061 nan 0.5470 -0.1501
## 2840 36.7004 nan 0.5470 -0.1380
## 2860 36.5675 nan 0.5470 -0.0801
## 2880 36.4309 nan 0.5470 -0.0847
## 2900 36.3214 nan 0.5470 -0.1583
## 2920 36.2667 nan 0.5470 -0.2150
## 2940 36.1818 nan 0.5470 -0.3765
## 2960 36.0877 nan 0.5470 -0.2005
## 2980 36.0490 nan 0.5470 -0.1911
## 3000 36.0528 nan 0.5470 -0.1971
## 3020 35.7958 nan 0.5470 -0.0586
## 3040 35.6981 nan 0.5470 -0.0549
## 3060 35.7219 nan 0.5470 -0.1138
## 3080 35.7903 nan 0.5470 -0.4219
## 3100 35.5469 nan 0.5470 -0.1295
## 3120 35.4998 nan 0.5470 -0.1575
## 3140 35.4864 nan 0.5470 0.0087
## 3160 35.3572 nan 0.5470 -0.0387
## 3180 35.2754 nan 0.5470 -0.0959
## 3200 35.2579 nan 0.5470 -0.1290
## 3220 35.2564 nan 0.5470 -0.1930
## 3240 35.0663 nan 0.5470 -0.0564
## 3260 35.0350 nan 0.5470 -0.2138
## 3280 34.8091 nan 0.5470 -0.1120
## 3300 34.7420 nan 0.5470 -0.1413
## 3320 34.7197 nan 0.5470 -0.1303
## 3340 34.7141 nan 0.5470 -0.3466
## 3360 34.7008 nan 0.5470 -0.1359
## 3380 34.5466 nan 0.5470 -0.1431
## 3400 34.4990 nan 0.5470 -0.0861
## 3420 34.4874 nan 0.5470 -0.0457
## 3440 34.3689 nan 0.5470 -0.1197
## 3460 34.2587 nan 0.5470 -0.1860
## 3480 34.1296 nan 0.5470 -0.1261
## 3489 34.0941 nan 0.5470 -0.0940
##
## Iter TrainDeviance ValidDeviance StepSize Improve
## 1 824.5621 nan 0.2306 357.5419
## 2 606.2890 nan 0.2306 218.9355
## 3 464.9577 nan 0.2306 137.8700
## 4 376.8895 nan 0.2306 88.1293
## 5 319.7340 nan 0.2306 57.5879
## 6 280.9460 nan 0.2306 37.6326
## 7 252.7670 nan 0.2306 27.6333
## 8 234.4730 nan 0.2306 17.5855
## 9 221.0968 nan 0.2306 13.1766
## 10 210.9191 nan 0.2306 9.2725
## 20 167.0790 nan 0.2306 1.4132
## 40 134.3664 nan 0.2306 0.3136
## 60 116.9811 nan 0.2306 0.2405
## 80 106.6758 nan 0.2306 -0.2214
## 100 97.7058 nan 0.2306 0.0582
## 120 90.6339 nan 0.2306 0.0202
## 140 85.1634 nan 0.2306 -0.1643
## 160 79.9563 nan 0.2306 -0.2670
## 180 75.9831 nan 0.2306 -0.0709
## 200 72.1967 nan 0.2306 -0.1579
## 220 69.0412 nan 0.2306 -0.1711
## 240 66.0974 nan 0.2306 -0.1587
## 260 63.4029 nan 0.2306 -0.1677
## 280 61.0143 nan 0.2306 -0.2201
## 300 58.8483 nan 0.2306 0.0683
## 320 56.6593 nan 0.2306 -0.0757
## 340 55.0994 nan 0.2306 -0.1701
## 360 53.6138 nan 0.2306 -0.1535
## 380 52.1406 nan 0.2306 -0.1320
## 400 50.8933 nan 0.2306 -0.3028
## 420 49.7309 nan 0.2306 -0.3120
## 440 48.5677 nan 0.2306 -0.0999
## 460 47.2341 nan 0.2306 -0.1240
## 480 46.1920 nan 0.2306 -0.2287
## 500 45.1822 nan 0.2306 -0.2430
## 520 44.2399 nan 0.2306 -0.0932
## 540 43.4788 nan 0.2306 -0.1985
## 560 42.7071 nan 0.2306 -0.1199
## 580 41.8763 nan 0.2306 -0.0849
## 600 41.1899 nan 0.2306 -0.1755
## 620 40.6568 nan 0.2306 -0.1077
## 640 40.1204 nan 0.2306 -0.1063
## 660 39.5524 nan 0.2306 -0.1604
## 680 38.9617 nan 0.2306 -0.1168
## 700 38.5263 nan 0.2306 -0.1890
## 720 38.0407 nan 0.2306 -0.1895
## 740 37.6491 nan 0.2306 -0.1075
## 760 37.1249 nan 0.2306 -0.1657
## 780 36.7749 nan 0.2306 -0.1646
## 800 36.3994 nan 0.2306 -0.1476
## 820 36.0038 nan 0.2306 -0.1727
## 840 35.5173 nan 0.2306 -0.1111
## 860 35.1277 nan 0.2306 -0.1104
## 880 34.7810 nan 0.2306 -0.1301
## 900 34.5154 nan 0.2306 -0.1456
## 920 34.2253 nan 0.2306 -0.1148
## 940 33.8886 nan 0.2306 -0.1617
## 960 33.6057 nan 0.2306 -0.1763
## 980 33.3641 nan 0.2306 -0.1493
## 1000 33.0285 nan 0.2306 -0.1065
## 1020 32.8366 nan 0.2306 -0.2730
## 1040 32.5874 nan 0.2306 -0.1524
## 1060 32.3549 nan 0.2306 -0.3181
## 1080 32.0719 nan 0.2306 -0.0423
## 1100 31.8083 nan 0.2306 -0.0673
## 1120 31.5794 nan 0.2306 -0.1368
## 1140 31.3509 nan 0.2306 -0.0799
## 1160 31.1488 nan 0.2306 -0.1068
## 1180 30.9390 nan 0.2306 -0.1364
## 1200 30.7420 nan 0.2306 -0.1169
## 1220 30.5889 nan 0.2306 -0.1313
## 1240 30.3853 nan 0.2306 -0.1398
## 1260 30.1303 nan 0.2306 -0.0919
## 1280 29.8632 nan 0.2306 -0.0543
## 1300 29.7769 nan 0.2306 -0.1612
## 1320 29.5437 nan 0.2306 -0.1324
## 1340 29.4203 nan 0.2306 -0.1609
## 1360 29.2314 nan 0.2306 -0.2391
## 1380 29.0911 nan 0.2306 -0.0561
## 1400 28.9397 nan 0.2306 -0.0836
## 1420 28.8238 nan 0.2306 -0.1213
## 1440 28.6886 nan 0.2306 -0.1308
## 1460 28.5297 nan 0.2306 -0.1295
## 1480 28.4300 nan 0.2306 -0.1966
## 1500 28.3322 nan 0.2306 -0.0946
## 1520 28.2422 nan 0.2306 -0.1913
## 1540 28.1231 nan 0.2306 -0.1030
## 1560 28.0336 nan 0.2306 -0.1264
## 1580 27.9447 nan 0.2306 -0.1251
## 1600 27.8523 nan 0.2306 -0.1244
## 1620 27.7223 nan 0.2306 -0.1346
## 1640 27.6466 nan 0.2306 -0.1866
## 1660 27.5731 nan 0.2306 -0.1287
## 1680 27.4824 nan 0.2306 -0.1574
## 1700 27.4106 nan 0.2306 -0.1153
## 1720 27.3141 nan 0.2306 -0.0674
## 1740 27.2428 nan 0.2306 -0.1385
## 1760 27.0648 nan 0.2306 -0.0825
## 1780 27.0419 nan 0.2306 -0.2381
## 1800 26.9406 nan 0.2306 -0.0875
## 1820 26.8727 nan 0.2306 -0.1115
## 1840 26.7248 nan 0.2306 -0.1537
## 1860 26.6582 nan 0.2306 -0.1143
## 1880 26.5882 nan 0.2306 -0.1182
## 1900 26.5242 nan 0.2306 -0.2009
## 1920 26.4862 nan 0.2306 -0.1157
## 1940 26.4280 nan 0.2306 -0.1249
## 1960 26.3105 nan 0.2306 -0.1185
## 1980 26.2337 nan 0.2306 -0.1599
## 2000 26.1559 nan 0.2306 -0.1213
## 2020 26.0984 nan 0.2306 -0.0990
## 2040 26.0258 nan 0.2306 -0.0645
## 2060 26.0332 nan 0.2306 -0.1599
## 2080 25.9343 nan 0.2306 -0.0734
## 2100 25.9254 nan 0.2306 -0.0652
## 2120 25.8465 nan 0.2306 -0.0960
## 2140 25.7692 nan 0.2306 -0.1480
## 2160 25.6890 nan 0.2306 -0.1098
## 2180 25.5939 nan 0.2306 -0.0725
## 2200 25.5302 nan 0.2306 -0.1502
## 2220 25.4763 nan 0.2306 -0.1986
## 2240 25.4023 nan 0.2306 -0.2125
## 2260 25.3471 nan 0.2306 -0.0712
## 2280 25.3249 nan 0.2306 -0.1573
## 2300 25.2497 nan 0.2306 -0.0913
## 2320 25.1908 nan 0.2306 -0.0971
## 2340 25.1392 nan 0.2306 -0.1003
## 2360 25.0913 nan 0.2306 -0.1520
## 2380 25.0346 nan 0.2306 -0.1765
## 2400 24.9952 nan 0.2306 -0.0902
## 2420 24.9476 nan 0.2306 -0.1496
## 2440 24.8756 nan 0.2306 -0.1015
## 2460 24.8935 nan 0.2306 -0.1575
## 2480 24.8325 nan 0.2306 -0.1369
## 2500 24.7295 nan 0.2306 -0.0833
## 2520 24.7552 nan 0.2306 -0.1138
## 2540 24.7335 nan 0.2306 -0.0892
## 2560 24.7007 nan 0.2306 -0.0727
## 2580 24.6701 nan 0.2306 -0.2264
## 2600 24.6286 nan 0.2306 -0.1302
## 2620 24.5795 nan 0.2306 -0.1299
## 2640 24.5283 nan 0.2306 -0.0933
## 2660 24.5215 nan 0.2306 -0.1407
## 2680 24.5017 nan 0.2306 -0.0754
## 2700 24.4857 nan 0.2306 -0.1643
## 2720 24.4955 nan 0.2306 -0.1050
## 2740 24.4636 nan 0.2306 -0.0847
## 2760 24.4040 nan 0.2306 -0.1042
## 2780 24.3654 nan 0.2306 -0.1121
## 2800 24.3338 nan 0.2306 -0.0876
## 2820 24.3491 nan 0.2306 -0.0877
## 2840 24.2922 nan 0.2306 -0.0985
## 2860 24.2357 nan 0.2306 -0.0984
## 2880 24.2008 nan 0.2306 -0.1560
## 2900 24.1793 nan 0.2306 -0.0503
## 2920 24.1016 nan 0.2306 -0.1304
## 2940 24.0673 nan 0.2306 -0.1723
## 2960 24.0149 nan 0.2306 -0.0551
## 2980 23.9990 nan 0.2306 -0.0836
## 3000 23.9951 nan 0.2306 -0.0907
## 3020 23.9694 nan 0.2306 -0.1882
## 3040 23.9402 nan 0.2306 -0.1880
## 3060 23.8905 nan 0.2306 -0.2435
## 3080 23.8815 nan 0.2306 -0.1768
## 3100 23.8178 nan 0.2306 -0.1400
## 3120 23.8653 nan 0.2306 -0.0735
## 3140 23.7836 nan 0.2306 -0.1773
## 3160 23.7492 nan 0.2306 -0.1066
## 3180 23.7274 nan 0.2306 -0.1466
## 3200 23.6721 nan 0.2306 -0.0547
## 3220 23.6558 nan 0.2306 -0.1184
## 3240 23.6412 nan 0.2306 -0.1136
## 3260 23.6023 nan 0.2306 -0.1368
## 3280 23.6091 nan 0.2306 -0.1070
## 3300 23.5638 nan 0.2306 -0.1044
## 3320 23.5612 nan 0.2306 -0.1122
## 3340 23.5538 nan 0.2306 -0.1314
## 3360 23.5532 nan 0.2306 -0.0703
## 3380 23.5061 nan 0.2306 -0.2079
## 3400 23.4527 nan 0.2306 -0.0993
## 3420 23.4089 nan 0.2306 -0.1353
## 3440 23.3772 nan 0.2306 -0.1417
## 3460 23.4300 nan 0.2306 -0.1320
## 3480 23.4011 nan 0.2306 -0.1086
## 3500 23.3875 nan 0.2306 -0.0988
## 3520 23.3403 nan 0.2306 -0.1481
## 3540 23.3376 nan 0.2306 -0.0600
## 3560 23.3049 nan 0.2306 -0.1296
## 3580 23.2585 nan 0.2306 -0.1017
## 3600 23.2468 nan 0.2306 -0.0890
## 3620 23.2244 nan 0.2306 -0.0580
## 3640 23.1967 nan 0.2306 -0.1625
## 3660 23.1718 nan 0.2306 -0.1127
## 3680 23.1465 nan 0.2306 -0.0660
## 3700 23.1517 nan 0.2306 -0.0973
## 3720 23.1168 nan 0.2306 -0.0579
##
## Iter TrainDeviance ValidDeviance StepSize Improve
## 1 672.2928 nan 0.3896 502.1240
## 2 468.8291 nan 0.3896 202.9958
## 3 373.2304 nan 0.3896 91.2630
## 4 325.7640 nan 0.3896 45.9424
## 5 300.9983 nan 0.3896 23.7617
## 6 285.5462 nan 0.3896 14.4377
## 7 270.9259 nan 0.3896 14.1772
## 8 262.2058 nan 0.3896 7.7425
## 9 254.9525 nan 0.3896 6.0039
## 10 244.5546 nan 0.3896 10.4672
## 20 204.2044 nan 0.3896 0.4214
## 40 171.9674 nan 0.3896 0.5796
## 60 155.5397 nan 0.3896 0.0694
## 80 141.6836 nan 0.3896 -0.0275
## 100 132.1458 nan 0.3896 -0.1046
## 120 124.7488 nan 0.3896 -0.0552
## 140 120.0868 nan 0.3896 -0.1835
## 160 115.1790 nan 0.3896 0.0835
## 180 110.1663 nan 0.3896 -0.1350
## 200 106.6357 nan 0.3896 0.1294
## 220 103.2750 nan 0.3896 -0.1403
## 240 101.0129 nan 0.3896 -0.3550
## 260 98.7296 nan 0.3896 -0.2488
## 280 95.5102 nan 0.3896 -0.0320
## 300 93.0493 nan 0.3896 -0.0371
## 320 90.9065 nan 0.3896 -0.0818
## 340 88.9585 nan 0.3896 -0.1094
## 360 86.9967 nan 0.3896 0.0113
## 380 85.0310 nan 0.3896 -0.1480
## 400 83.3119 nan 0.3896 -0.3268
## 420 82.0764 nan 0.3896 -0.0950
## 440 80.5235 nan 0.3896 -0.1702
## 460 79.0738 nan 0.3896 -0.0765
## 480 77.8660 nan 0.3896 -0.0070
## 500 76.7659 nan 0.3896 -0.1939
## 520 75.4761 nan 0.3896 -0.1132
## 540 74.2228 nan 0.3896 -0.2946
## 560 73.1614 nan 0.3896 -0.0490
## 580 72.3256 nan 0.3896 -0.0422
## 600 71.3000 nan 0.3896 -0.1747
## 620 70.4333 nan 0.3896 -0.2136
## 640 69.2329 nan 0.3896 -0.0623
## 660 68.2587 nan 0.3896 -0.0844
## 680 67.2746 nan 0.3896 -0.2405
## 700 66.4487 nan 0.3896 -0.1245
## 720 65.4298 nan 0.3896 -0.1606
## 740 64.6610 nan 0.3896 -0.0281
## 760 63.7985 nan 0.3896 -0.1438
## 780 62.9984 nan 0.3896 -0.1443
## 800 62.3878 nan 0.3896 -0.1029
## 820 61.6125 nan 0.3896 -0.0702
## 840 60.7782 nan 0.3896 -0.1082
## 860 60.1579 nan 0.3896 -0.1305
## 880 59.5151 nan 0.3896 -0.1190
## 900 58.9498 nan 0.3896 -0.1683
## 920 58.4945 nan 0.3896 -0.1848
## 940 57.9744 nan 0.3896 -0.1616
## 960 57.5913 nan 0.3896 -0.1971
## 980 57.0559 nan 0.3896 -0.0654
## 1000 56.5428 nan 0.3896 -0.0835
## 1020 56.0524 nan 0.3896 -0.1548
## 1040 55.7051 nan 0.3896 -0.1439
## 1060 55.2176 nan 0.3896 -0.0929
## 1080 54.8356 nan 0.3896 -0.0817
## 1100 54.5203 nan 0.3896 -0.1427
## 1120 54.1719 nan 0.3896 -0.0524
## 1140 53.6079 nan 0.3896 -0.1075
## 1160 53.2811 nan 0.3896 -0.0482
## 1180 52.9364 nan 0.3896 -0.0517
## 1200 52.5529 nan 0.3896 -0.0878
## 1220 52.4031 nan 0.3896 -0.1083
## 1240 52.1374 nan 0.3896 -0.1739
## 1260 51.7621 nan 0.3896 -0.1317
## 1280 51.4563 nan 0.3896 -0.1591
## 1300 51.1539 nan 0.3896 -0.1190
## 1320 50.8621 nan 0.3896 -0.2975
## 1340 50.4855 nan 0.3896 -0.1398
## 1360 50.1061 nan 0.3896 -0.1451
## 1380 49.9010 nan 0.3896 -0.1168
## 1400 49.5773 nan 0.3896 -0.0791
## 1420 49.2619 nan 0.3896 -0.0283
## 1440 48.9560 nan 0.3896 -0.1042
## 1460 48.6519 nan 0.3896 -0.1079
## 1480 48.3948 nan 0.3896 -0.0216
## 1500 48.1869 nan 0.3896 -0.1422
## 1520 47.8576 nan 0.3896 -0.0621
## 1540 47.5032 nan 0.3896 -0.1407
## 1560 47.2837 nan 0.3896 -0.0434
## 1580 47.0235 nan 0.3896 -0.0578
## 1600 46.8355 nan 0.3896 -0.1391
## 1620 46.5850 nan 0.3896 -0.0107
## 1640 46.2851 nan 0.3896 -0.0727
## 1660 46.0476 nan 0.3896 -0.0829
## 1680 45.8815 nan 0.3896 -0.1056
## 1700 45.6395 nan 0.3896 -0.0751
## 1720 45.4020 nan 0.3896 -0.1121
## 1740 45.1187 nan 0.3896 -0.0631
## 1760 44.9604 nan 0.3896 -0.0813
## 1780 44.8103 nan 0.3896 -0.1420
## 1800 44.7552 nan 0.3896 -0.1089
## 1820 44.5283 nan 0.3896 -0.1610
## 1840 44.3517 nan 0.3896 -0.0852
## 1860 44.0921 nan 0.3896 -0.1286
## 1880 43.8645 nan 0.3896 -0.0635
## 1900 43.6592 nan 0.3896 -0.0743
## 1920 43.4406 nan 0.3896 -0.1241
## 1940 43.3249 nan 0.3896 -0.1290
## 1960 43.1072 nan 0.3896 -0.0964
## 1980 42.9389 nan 0.3896 -0.1754
## 2000 42.7217 nan 0.3896 -0.0554
## 2020 42.5560 nan 0.3896 -0.1556
## 2040 42.3197 nan 0.3896 -0.1205
## 2060 42.1095 nan 0.3896 -0.0815
## 2080 42.1211 nan 0.3896 -0.4315
## 2100 41.8606 nan 0.3896 -0.1479
## 2120 41.6666 nan 0.3896 -0.0858
## 2140 41.5360 nan 0.3896 -0.1198
## 2160 41.4672 nan 0.3896 -0.1682
## 2180 41.2682 nan 0.3896 -0.1166
## 2200 41.0692 nan 0.3896 -0.0673
## 2220 40.8680 nan 0.3896 -0.0759
## 2240 40.6845 nan 0.3896 -0.1000
## 2260 40.5514 nan 0.3896 -0.0721
## 2280 40.4866 nan 0.3896 -0.0714
## 2300 40.3256 nan 0.3896 -0.0852
## 2320 40.1994 nan 0.3896 -0.0645
## 2340 40.0432 nan 0.3896 -0.0780
## 2360 39.8786 nan 0.3896 -0.0376
## 2380 39.7998 nan 0.3896 -0.1558
## 2400 39.6718 nan 0.3896 -0.2468
## 2420 39.5642 nan 0.3896 -0.0125
## 2440 39.4364 nan 0.3896 -0.1086
## 2460 39.3158 nan 0.3896 -0.0223
## 2480 39.2725 nan 0.3896 -0.1183
## 2500 39.0910 nan 0.3896 -0.0667
## 2520 38.9101 nan 0.3896 -0.0691
## 2540 38.7957 nan 0.3896 -0.0745
## 2560 38.6868 nan 0.3896 -0.0642
## 2580 38.4327 nan 0.3896 -0.1084
## 2600 38.3856 nan 0.3896 -0.1249
## 2620 38.3827 nan 0.3896 -0.1012
## 2640 38.1964 nan 0.3896 -0.0952
## 2660 38.1900 nan 0.3896 -0.1800
## 2680 37.9421 nan 0.3896 -0.0780
## 2700 37.8584 nan 0.3896 -0.1163
## 2720 37.7215 nan 0.3896 -0.1507
## 2740 37.5946 nan 0.3896 -0.0467
## 2760 37.5014 nan 0.3896 -0.0576
## 2780 37.4265 nan 0.3896 -0.0753
## 2800 37.2722 nan 0.3896 -0.0368
## 2820 37.1668 nan 0.3896 -0.0403
## 2840 37.1104 nan 0.3896 -0.1121
## 2860 37.0961 nan 0.3896 -0.0977
## 2880 37.0555 nan 0.3896 -0.0575
## 2900 36.8989 nan 0.3896 -0.1192
## 2920 36.7971 nan 0.3896 -0.1223
## 2940 36.7313 nan 0.3896 -0.0757
## 2960 36.5919 nan 0.3896 -0.1070
## 2980 36.5567 nan 0.3896 -0.0918
## 3000 36.4663 nan 0.3896 -0.0495
## 3020 36.3946 nan 0.3896 -0.0797
## 3040 36.3119 nan 0.3896 -0.0674
## 3060 36.2611 nan 0.3896 -0.1382
## 3080 36.1556 nan 0.3896 -0.0164
## 3100 36.0812 nan 0.3896 -0.1237
## 3120 35.9914 nan 0.3896 -0.0693
## 3140 35.9102 nan 0.3896 -0.0753
## 3160 35.8598 nan 0.3896 -0.1289
## 3180 35.7090 nan 0.3896 -0.0674
## 3200 35.6170 nan 0.3896 -0.0515
## 3220 35.5703 nan 0.3896 -0.0703
## 3240 35.4779 nan 0.3896 -0.1079
## 3260 35.4500 nan 0.3896 -0.1812
## 3280 35.4186 nan 0.3896 -0.1514
## 3300 35.3465 nan 0.3896 -0.0478
## 3320 35.2811 nan 0.3896 -0.1351
## 3340 35.2310 nan 0.3896 -0.1079
## 3360 35.1391 nan 0.3896 -0.0602
## 3380 35.0956 nan 0.3896 -0.1036
## 3400 35.0512 nan 0.3896 -0.1380
## 3420 34.9733 nan 0.3896 -0.1474
## 3440 34.8654 nan 0.3896 -0.0503
## 3460 34.7549 nan 0.3896 -0.0301
## 3480 34.6856 nan 0.3896 -0.0473
## 3500 34.6568 nan 0.3896 -0.0720
## 3520 34.6509 nan 0.3896 -0.0805
## 3540 34.5544 nan 0.3896 -0.0728
## 3560 34.4874 nan 0.3896 -0.0901
## 3580 34.4388 nan 0.3896 -0.0789
## 3600 34.3650 nan 0.3896 -0.0359
## 3620 34.2992 nan 0.3896 -0.0524
## 3640 34.2083 nan 0.3896 -0.1112
## 3660 34.1309 nan 0.3896 -0.0811
## 3680 34.0759 nan 0.3896 -0.0659
## 3700 33.9991 nan 0.3896 -0.0599
## 3720 34.0195 nan 0.3896 -0.1157
## 3740 33.9021 nan 0.3896 -0.1010
## 3760 33.8605 nan 0.3896 -0.0676
## 3780 33.8399 nan 0.3896 -0.3372
## 3800 33.6721 nan 0.3896 -0.0778
## 3820 33.6651 nan 0.3896 -0.0773
## 3840 33.5578 nan 0.3896 -0.0943
## 3860 33.5280 nan 0.3896 -0.0809
## 3880 33.4816 nan 0.3896 -0.0736
## 3900 33.4687 nan 0.3896 -0.0832
## 3920 33.3714 nan 0.3896 -0.1173
## 3940 33.3102 nan 0.3896 -0.1358
## 3960 33.2368 nan 0.3896 -0.1130
## 3980 33.2219 nan 0.3896 -0.1469
## 4000 33.1143 nan 0.3896 -0.0863
## 4020 33.1301 nan 0.3896 -0.1614
## 4040 33.0614 nan 0.3896 -0.0709
## 4060 32.9756 nan 0.3896 -0.0843
## 4080 32.9398 nan 0.3896 -0.0840
##
## Iter TrainDeviance ValidDeviance StepSize Improve
## 1 546.2948 nan 0.5470 646.6852
## 2 378.2894 nan 0.5470 163.8886
## 3 321.7950 nan 0.5470 52.8618
## 4 293.7752 nan 0.5470 25.7071
## 5 278.4390 nan 0.5470 15.5100
## 6 266.7310 nan 0.5470 11.0675
## 7 255.6124 nan 0.5470 9.3796
## 8 245.9038 nan 0.5470 9.1618
## 9 239.9977 nan 0.5470 4.6418
## 10 235.6126 nan 0.5470 3.2009
## 20 199.2822 nan 0.5470 1.2231
## 40 163.1671 nan 0.5470 -0.0040
## 60 148.7943 nan 0.5470 -0.2149
## 80 138.7075 nan 0.5470 -0.4525
## 100 129.6375 nan 0.5470 -0.0188
## 120 122.2797 nan 0.5470 -0.3828
## 140 115.4769 nan 0.5470 -0.0544
## 160 109.6355 nan 0.5470 -0.3433
## 180 105.3269 nan 0.5470 -0.3373
## 200 102.2365 nan 0.5470 -0.3365
## 220 99.4703 nan 0.5470 -0.1588
## 240 96.4003 nan 0.5470 -0.2000
## 260 94.0561 nan 0.5470 -0.4103
## 280 91.2289 nan 0.5470 -0.3090
## 300 88.4029 nan 0.5470 -0.2524
## 320 86.3056 nan 0.5470 -0.1320
## 340 83.9285 nan 0.5470 -0.4618
## 360 82.3247 nan 0.5470 -0.2013
## 380 81.1313 nan 0.5470 -0.1909
## 400 78.9120 nan 0.5470 -0.2479
## 420 77.2574 nan 0.5470 -0.1592
## 440 75.6182 nan 0.5470 -0.1872
## 460 74.3703 nan 0.5470 -0.3239
## 480 73.1834 nan 0.5470 -0.2688
## 500 72.1455 nan 0.5470 -0.2169
## 520 71.0772 nan 0.5470 -0.3412
## 540 70.2103 nan 0.5470 -0.1895
## 560 69.3643 nan 0.5470 -0.1231
## 580 68.2779 nan 0.5470 -0.2247
## 600 67.4116 nan 0.5470 -0.1835
## 620 66.7119 nan 0.5470 -0.0942
## 640 65.8136 nan 0.5470 -0.1245
## 660 64.9947 nan 0.5470 -0.2303
## 680 64.1698 nan 0.5470 -0.2024
## 700 63.3151 nan 0.5470 -0.2006
## 720 62.7323 nan 0.5470 -0.1402
## 740 62.0844 nan 0.5470 -0.1707
## 760 61.6574 nan 0.5470 -0.1431
## 780 60.9029 nan 0.5470 -0.0909
## 800 60.3089 nan 0.5470 -0.2176
## 820 59.4568 nan 0.5470 0.0028
## 840 58.6875 nan 0.5470 -0.1739
## 860 58.3116 nan 0.5470 -0.1101
## 880 57.5041 nan 0.5470 -0.1423
## 900 57.1253 nan 0.5470 -0.1580
## 920 56.5548 nan 0.5470 -0.1564
## 940 56.2386 nan 0.5470 -0.2200
## 960 55.6240 nan 0.5470 -0.1016
## 980 55.0811 nan 0.5470 -0.0577
## 1000 54.6083 nan 0.5470 -0.1289
## 1020 54.2450 nan 0.5470 -0.1850
## 1040 53.9240 nan 0.5470 -0.2109
## 1060 53.4811 nan 0.5470 -0.0486
## 1080 53.2204 nan 0.5470 -0.1069
## 1100 52.8317 nan 0.5470 -0.2187
## 1120 52.3022 nan 0.5470 -0.1560
## 1140 52.0151 nan 0.5470 -0.3351
## 1160 51.5292 nan 0.5470 -0.2224
## 1180 51.1206 nan 0.5470 -0.1138
## 1200 50.7515 nan 0.5470 -0.0862
## 1220 50.4533 nan 0.5470 -0.0963
## 1240 50.2646 nan 0.5470 -0.1898
## 1260 49.9070 nan 0.5470 -0.2007
## 1280 49.4419 nan 0.5470 -0.1435
## 1300 49.0973 nan 0.5470 -0.1621
## 1320 48.7511 nan 0.5470 -0.2605
## 1340 48.4590 nan 0.5470 -0.1636
## 1360 48.0815 nan 0.5470 -0.1647
## 1380 47.9041 nan 0.5470 -0.3045
## 1400 47.8782 nan 0.5470 -0.3376
## 1420 47.5136 nan 0.5470 -0.3805
## 1440 47.2475 nan 0.5470 -0.2492
## 1460 47.0586 nan 0.5470 -0.4874
## 1480 46.9318 nan 0.5470 -0.2200
## 1500 46.7240 nan 0.5470 -0.0487
## 1520 46.4873 nan 0.5470 -0.2823
## 1540 46.3077 nan 0.5470 -0.2648
## 1560 46.0930 nan 0.5470 -0.0698
## 1580 45.8568 nan 0.5470 -0.1801
## 1600 45.5679 nan 0.5470 -0.2215
## 1620 45.3332 nan 0.5470 -0.0180
## 1640 45.2090 nan 0.5470 -0.0639
## 1660 45.1088 nan 0.5470 -0.2155
## 1680 44.8143 nan 0.5470 -0.1830
## 1700 44.6326 nan 0.5470 -0.3710
## 1720 44.5032 nan 0.5470 -0.2767
## 1740 44.2053 nan 0.5470 -0.1053
## 1760 44.1570 nan 0.5470 -0.2717
## 1780 43.9184 nan 0.5470 -0.1711
## 1800 43.6966 nan 0.5470 -0.1875
## 1820 43.5577 nan 0.5470 -0.1331
## 1840 43.3955 nan 0.5470 -0.1207
## 1860 43.2338 nan 0.5470 -0.1692
## 1880 42.9989 nan 0.5470 -0.1791
## 1900 42.8895 nan 0.5470 -0.1618
## 1920 42.7325 nan 0.5470 -0.1135
## 1940 42.5153 nan 0.5470 -0.0952
## 1960 42.4507 nan 0.5470 -0.0847
## 1980 42.4300 nan 0.5470 -0.1617
## 2000 42.2852 nan 0.5470 -0.1925
## 2020 42.1706 nan 0.5470 -0.1224
## 2040 42.0260 nan 0.5470 -0.0835
## 2060 41.8698 nan 0.5470 -0.1809
## 2080 41.7229 nan 0.5470 -0.1633
## 2100 41.3905 nan 0.5470 -0.2854
## 2120 41.3040 nan 0.5470 -0.1320
## 2140 41.1448 nan 0.5470 -0.1122
## 2160 41.0695 nan 0.5470 -0.1080
## 2180 40.8824 nan 0.5470 -0.1621
## 2200 40.6822 nan 0.5470 -0.0863
## 2220 40.6808 nan 0.5470 -0.1476
## 2240 40.5806 nan 0.5470 -0.1180
## 2260 40.3749 nan 0.5470 -0.1889
## 2280 40.3416 nan 0.5470 -0.2041
## 2300 40.3199 nan 0.5470 -0.2025
## 2320 40.0482 nan 0.5470 -0.1703
## 2340 39.8630 nan 0.5470 -0.1089
## 2360 39.7439 nan 0.5470 -0.2013
## 2380 39.7414 nan 0.5470 -0.3464
## 2400 39.5554 nan 0.5470 -0.0857
## 2420 39.4500 nan 0.5470 -0.0765
## 2440 39.3318 nan 0.5470 -0.0608
## 2460 39.2688 nan 0.5470 -0.2066
## 2480 39.1404 nan 0.5470 -0.1837
## 2500 39.0650 nan 0.5470 -0.1807
## 2520 39.0264 nan 0.5470 -0.2322
## 2540 38.8263 nan 0.5470 -0.0932
## 2560 38.6978 nan 0.5470 -0.1369
## 2580 38.6525 nan 0.5470 -0.0984
## 2600 38.5024 nan 0.5470 -0.2570
## 2620 38.3330 nan 0.5470 -0.1617
## 2640 38.3354 nan 0.5470 -0.1706
## 2660 38.2057 nan 0.5470 -0.2021
## 2680 38.1418 nan 0.5470 -0.3236
## 2700 38.0967 nan 0.5470 -0.1170
## 2720 38.0542 nan 0.5470 -0.1910
## 2740 37.9633 nan 0.5470 -0.0686
## 2760 37.8519 nan 0.5470 -0.1214
## 2780 37.8009 nan 0.5470 -0.1629
## 2800 37.5930 nan 0.5470 -0.1508
## 2820 37.5071 nan 0.5470 -0.1648
## 2840 37.3790 nan 0.5470 -0.1243
## 2860 37.3053 nan 0.5470 -0.3473
## 2880 37.1731 nan 0.5470 -0.0932
## 2900 37.1060 nan 0.5470 -0.0827
## 2920 37.0209 nan 0.5470 -0.2038
## 2940 36.9293 nan 0.5470 -0.1604
## 2960 36.8635 nan 0.5470 -0.1156
## 2980 36.7335 nan 0.5470 -0.1661
## 3000 36.6274 nan 0.5470 -0.1797
## 3020 36.5727 nan 0.5470 -0.1609
## 3040 36.4889 nan 0.5470 -0.3193
## 3060 36.3668 nan 0.5470 -0.1462
## 3080 36.3386 nan 0.5470 -0.1691
## 3100 36.2291 nan 0.5470 -0.1268
## 3120 36.1105 nan 0.5470 -0.0544
## 3140 36.0828 nan 0.5470 -0.0783
## 3160 36.0685 nan 0.5470 -0.0893
## 3180 35.9234 nan 0.5470 -0.2233
## 3200 35.8534 nan 0.5470 -0.0961
## 3220 35.7648 nan 0.5470 -0.1794
## 3240 35.7871 nan 0.5470 -0.0977
## 3260 35.6705 nan 0.5470 -0.1913
## 3280 35.6304 nan 0.5470 -0.2171
## 3300 35.5863 nan 0.5470 -0.0343
## 3320 35.5213 nan 0.5470 -0.3126
## 3340 35.3469 nan 0.5470 -0.0598
## 3360 35.2604 nan 0.5470 -0.0945
## 3380 35.2411 nan 0.5470 -0.0335
## 3400 35.1440 nan 0.5470 -0.1289
## 3420 35.1051 nan 0.5470 -0.1339
## 3440 34.9862 nan 0.5470 -0.1835
## 3460 34.9795 nan 0.5470 -0.1039
## 3480 34.9225 nan 0.5470 -0.2075
## 3489 34.8666 nan 0.5470 -0.0450
##
## Iter TrainDeviance ValidDeviance StepSize Improve
## 1 820.8455 nan 0.2306 354.9822
## 2 601.3447 nan 0.2306 221.3719
## 3 462.4436 nan 0.2306 142.1719
## 4 372.0282 nan 0.2306 88.0553
## 5 316.0236 nan 0.2306 54.6381
## 6 276.6129 nan 0.2306 38.8663
## 7 249.8146 nan 0.2306 25.2368
## 8 230.2540 nan 0.2306 18.0238
## 9 215.7105 nan 0.2306 13.2429
## 10 205.2733 nan 0.2306 10.4246
## 20 162.6827 nan 0.2306 1.6512
## 40 130.8305 nan 0.2306 0.1719
## 60 115.2979 nan 0.2306 0.2337
## 80 105.0743 nan 0.2306 0.0607
## 100 95.6104 nan 0.2306 0.0922
## 120 88.4424 nan 0.2306 0.0678
## 140 82.6230 nan 0.2306 0.0993
## 160 77.4984 nan 0.2306 -0.2168
## 180 73.1477 nan 0.2306 0.0106
## 200 69.3653 nan 0.2306 -0.2684
## 220 66.1511 nan 0.2306 -0.0490
## 240 63.4546 nan 0.2306 -0.1301
## 260 61.0822 nan 0.2306 -0.1345
## 280 58.8076 nan 0.2306 -0.0783
## 300 56.8362 nan 0.2306 -0.1391
## 320 55.2592 nan 0.2306 -0.2477
## 340 53.6932 nan 0.2306 -0.1620
## 360 52.1652 nan 0.2306 -0.2156
## 380 50.6490 nan 0.2306 -0.1580
## 400 49.2110 nan 0.2306 -0.2440
## 420 47.9137 nan 0.2306 -0.2674
## 440 46.6894 nan 0.2306 -0.1133
## 460 45.6235 nan 0.2306 -0.1626
## 480 44.6172 nan 0.2306 -0.0637
## 500 43.6384 nan 0.2306 -0.1232
## 520 42.9310 nan 0.2306 -0.1872
## 540 42.0784 nan 0.2306 -0.2099
## 560 41.3450 nan 0.2306 -0.1478
## 580 40.5073 nan 0.2306 -0.2028
## 600 39.7866 nan 0.2306 -0.1737
## 620 39.1431 nan 0.2306 -0.1512
## 640 38.3727 nan 0.2306 -0.0302
## 660 37.8539 nan 0.2306 -0.0958
## 680 37.3364 nan 0.2306 -0.0593
## 700 36.7116 nan 0.2306 -0.1116
## 720 36.1723 nan 0.2306 -0.1101
## 740 35.5622 nan 0.2306 -0.0697
## 760 35.2239 nan 0.2306 -0.1248
## 780 34.7138 nan 0.2306 -0.0543
## 800 34.3137 nan 0.2306 -0.0901
## 820 33.9592 nan 0.2306 -0.1319
## 840 33.5209 nan 0.2306 -0.1007
## 860 33.2034 nan 0.2306 -0.1304
## 880 32.9071 nan 0.2306 -0.1025
## 900 32.5408 nan 0.2306 -0.1662
## 920 32.2455 nan 0.2306 -0.1753
## 940 31.9424 nan 0.2306 -0.1323
## 960 31.6492 nan 0.2306 -0.0516
## 980 31.3611 nan 0.2306 -0.2308
## 1000 31.1184 nan 0.2306 -0.1357
## 1020 30.9322 nan 0.2306 -0.1751
## 1040 30.6770 nan 0.2306 -0.1206
## 1060 30.3561 nan 0.2306 -0.1079
## 1080 30.2245 nan 0.2306 -0.0943
## 1100 29.9961 nan 0.2306 -0.1016
## 1120 29.8451 nan 0.2306 -0.1242
## 1140 29.6636 nan 0.2306 -0.2337
## 1160 29.4589 nan 0.2306 -0.1522
## 1180 29.2674 nan 0.2306 -0.0378
## 1200 29.1133 nan 0.2306 -0.0541
## 1220 28.9750 nan 0.2306 -0.0804
## 1240 28.7649 nan 0.2306 -0.0546
## 1260 28.5861 nan 0.2306 -0.1158
## 1280 28.4401 nan 0.2306 -0.1958
## 1300 28.2575 nan 0.2306 -0.0755
## 1320 28.1019 nan 0.2306 -0.1030
## 1340 28.0197 nan 0.2306 -0.0727
## 1360 27.8332 nan 0.2306 -0.0924
## 1380 27.6418 nan 0.2306 -0.1491
## 1400 27.5299 nan 0.2306 -0.2789
## 1420 27.3899 nan 0.2306 -0.1038
## 1440 27.2938 nan 0.2306 -0.1312
## 1460 27.1269 nan 0.2306 -0.0834
## 1480 27.0086 nan 0.2306 -0.1189
## 1500 26.9040 nan 0.2306 -0.1318
## 1520 26.7718 nan 0.2306 -0.1507
## 1540 26.6671 nan 0.2306 -0.1589
## 1560 26.6276 nan 0.2306 -0.1045
## 1580 26.5524 nan 0.2306 -0.0850
## 1600 26.4420 nan 0.2306 -0.0789
## 1620 26.3800 nan 0.2306 -0.0967
## 1640 26.3203 nan 0.2306 -0.0728
## 1660 26.2180 nan 0.2306 -0.1067
## 1680 26.1285 nan 0.2306 -0.0919
## 1700 26.0193 nan 0.2306 -0.0876
## 1720 25.9382 nan 0.2306 -0.2086
## 1740 25.9106 nan 0.2306 -0.0943
## 1760 25.8118 nan 0.2306 -0.2383
## 1780 25.7735 nan 0.2306 -0.1156
## 1800 25.6633 nan 0.2306 -0.0781
## 1820 25.6001 nan 0.2306 -0.1508
## 1840 25.5471 nan 0.2306 -0.2581
## 1860 25.5071 nan 0.2306 -0.1536
## 1880 25.4010 nan 0.2306 -0.1116
## 1900 25.3121 nan 0.2306 -0.1107
## 1920 25.2249 nan 0.2306 -0.1343
## 1940 25.1992 nan 0.2306 -0.1216
## 1960 25.1093 nan 0.2306 -0.1900
## 1980 25.0284 nan 0.2306 -0.1138
## 2000 24.9462 nan 0.2306 -0.1025
## 2020 24.8520 nan 0.2306 -0.0827
## 2040 24.8091 nan 0.2306 -0.1176
## 2060 24.7543 nan 0.2306 -0.1729
## 2080 24.6769 nan 0.2306 -0.1087
## 2100 24.5993 nan 0.2306 -0.1143
## 2120 24.5210 nan 0.2306 -0.0930
## 2140 24.5088 nan 0.2306 -0.1709
## 2160 24.4415 nan 0.2306 -0.0719
## 2180 24.4091 nan 0.2306 -0.1482
## 2200 24.3308 nan 0.2306 -0.0760
## 2220 24.3125 nan 0.2306 -0.0832
## 2240 24.2587 nan 0.2306 -0.1730
## 2260 24.2549 nan 0.2306 -0.1995
## 2280 24.2311 nan 0.2306 -0.4088
## 2300 24.1390 nan 0.2306 -0.1089
## 2320 24.1055 nan 0.2306 -0.0811
## 2340 24.0185 nan 0.2306 -0.0769
## 2360 23.9985 nan 0.2306 -0.0715
## 2380 23.9474 nan 0.2306 -0.0760
## 2400 23.8886 nan 0.2306 -0.1033
## 2420 23.8287 nan 0.2306 -0.0929
## 2440 23.8317 nan 0.2306 -0.0867
## 2460 23.7180 nan 0.2306 -0.0782
## 2480 23.7171 nan 0.2306 -0.1096
## 2500 23.7327 nan 0.2306 -0.1386
## 2520 23.6725 nan 0.2306 -0.0839
## 2540 23.6641 nan 0.2306 -0.1258
## 2560 23.6380 nan 0.2306 -0.1406
## 2580 23.5763 nan 0.2306 -0.0818
## 2600 23.4928 nan 0.2306 -0.1596
## 2620 23.4754 nan 0.2306 -0.1141
## 2640 23.4244 nan 0.2306 -0.1122
## 2660 23.4068 nan 0.2306 -0.0707
## 2680 23.3700 nan 0.2306 -0.1072
## 2700 23.3937 nan 0.2306 -0.0594
## 2720 23.3372 nan 0.2306 -0.0901
## 2740 23.2851 nan 0.2306 -0.0870
## 2760 23.2509 nan 0.2306 -0.1403
## 2780 23.2248 nan 0.2306 -0.1162
## 2800 23.2223 nan 0.2306 -0.0830
## 2820 23.1378 nan 0.2306 -0.1100
## 2840 23.1477 nan 0.2306 -0.0739
## 2860 23.1184 nan 0.2306 -0.0443
## 2880 23.0516 nan 0.2306 -0.0942
## 2900 23.0343 nan 0.2306 -0.1441
## 2920 22.9955 nan 0.2306 -0.0795
## 2940 22.9627 nan 0.2306 -0.0969
## 2960 22.9186 nan 0.2306 -0.1463
## 2980 22.9171 nan 0.2306 -0.0503
## 3000 22.8413 nan 0.2306 -0.1440
## 3020 22.8432 nan 0.2306 -0.0830
## 3040 22.8031 nan 0.2306 -0.0561
## 3060 22.7793 nan 0.2306 -0.1274
## 3080 22.7646 nan 0.2306 -0.0808
## 3100 22.7344 nan 0.2306 -0.1910
## 3120 22.7732 nan 0.2306 -0.1088
## 3140 22.7272 nan 0.2306 -0.1098
## 3160 22.6877 nan 0.2306 -0.0581
## 3180 22.6992 nan 0.2306 -0.0559
## 3200 22.6681 nan 0.2306 -0.0425
## 3220 22.6572 nan 0.2306 -0.0952
## 3240 22.6080 nan 0.2306 -0.1073
## 3260 22.5732 nan 0.2306 -0.1589
## 3280 22.5537 nan 0.2306 -0.0945
## 3300 22.5530 nan 0.2306 -0.0704
## 3320 22.5272 nan 0.2306 -0.1838
## 3340 22.5175 nan 0.2306 -0.1509
## 3360 22.5171 nan 0.2306 -0.1300
## 3380 22.5199 nan 0.2306 -0.1240
## 3400 22.4720 nan 0.2306 -0.1304
## 3420 22.4254 nan 0.2306 -0.0897
## 3440 22.3961 nan 0.2306 -0.0845
## 3460 22.3754 nan 0.2306 -0.1086
## 3480 22.3219 nan 0.2306 -0.1136
## 3500 22.3381 nan 0.2306 -0.1780
## 3520 22.3273 nan 0.2306 -0.1204
## 3540 22.3090 nan 0.2306 -0.1400
## 3560 22.2884 nan 0.2306 -0.0890
## 3580 22.2583 nan 0.2306 -0.1676
## 3600 22.2299 nan 0.2306 -0.0843
## 3620 22.2379 nan 0.2306 -0.2166
## 3640 22.2244 nan 0.2306 -0.1199
## 3660 22.1460 nan 0.2306 -0.0753
## 3680 22.1563 nan 0.2306 -0.0849
## 3700 22.1860 nan 0.2306 -0.1065
## 3720 22.1490 nan 0.2306 -0.1010
##
## Iter TrainDeviance ValidDeviance StepSize Improve
## 1 676.0789 nan 0.3896 499.9389
## 2 457.5824 nan 0.3896 209.9930
## 3 367.0105 nan 0.3896 89.8997
## 4 317.9868 nan 0.3896 50.2982
## 5 288.4089 nan 0.3896 28.2877
## 6 272.0808 nan 0.3896 16.1806
## 7 260.7543 nan 0.3896 9.5682
## 8 252.5014 nan 0.3896 8.3156
## 9 244.5863 nan 0.3896 7.3241
## 10 239.1197 nan 0.3896 4.5480
## 20 199.4454 nan 0.3896 1.7038
## 40 168.6189 nan 0.3896 0.7368
## 60 150.9381 nan 0.3896 -0.2519
## 80 139.7138 nan 0.3896 -0.2871
## 100 131.0291 nan 0.3896 -0.0364
## 120 124.1422 nan 0.3896 0.2185
## 140 119.0259 nan 0.3896 -0.1551
## 160 114.4388 nan 0.3896 0.0591
## 180 110.4909 nan 0.3896 -0.0368
## 200 105.7577 nan 0.3896 -0.0705
## 220 102.5183 nan 0.3896 0.0297
## 240 99.0502 nan 0.3896 0.0340
## 260 96.6216 nan 0.3896 -0.1598
## 280 94.0879 nan 0.3896 -0.0930
## 300 91.5863 nan 0.3896 -0.0534
## 320 89.3238 nan 0.3896 -0.1809
## 340 86.9851 nan 0.3896 -0.1171
## 360 85.1070 nan 0.3896 -0.2688
## 380 83.1372 nan 0.3896 -0.0468
## 400 81.5594 nan 0.3896 -0.1133
## 420 80.0558 nan 0.3896 -0.1269
## 440 78.7844 nan 0.3896 -0.1188
## 460 77.5261 nan 0.3896 -0.1940
## 480 76.2224 nan 0.3896 -0.1700
## 500 74.7569 nan 0.3896 -0.1670
## 520 73.2339 nan 0.3896 -0.1225
## 540 71.9614 nan 0.3896 -0.1706
## 560 70.6810 nan 0.3896 -0.1161
## 580 69.6940 nan 0.3896 -0.2345
## 600 68.5283 nan 0.3896 -0.0489
## 620 67.5954 nan 0.3896 -0.1447
## 640 66.5349 nan 0.3896 -0.1046
## 660 65.6624 nan 0.3896 -0.2035
## 680 64.8135 nan 0.3896 -0.1296
## 700 64.0356 nan 0.3896 -0.1768
## 720 63.1804 nan 0.3896 -0.1331
## 740 62.2803 nan 0.3896 -0.1011
## 760 61.5865 nan 0.3896 -0.0284
## 780 60.9275 nan 0.3896 -0.1362
## 800 60.3281 nan 0.3896 -0.1209
## 820 59.7245 nan 0.3896 -0.0888
## 840 59.1819 nan 0.3896 -0.0956
## 860 58.5279 nan 0.3896 -0.0733
## 880 57.9908 nan 0.3896 -0.0488
## 900 57.5268 nan 0.3896 -0.1082
## 920 56.9183 nan 0.3896 -0.1119
## 940 56.3394 nan 0.3896 -0.1014
## 960 55.9408 nan 0.3896 -0.0925
## 980 55.3586 nan 0.3896 -0.1078
## 1000 54.6689 nan 0.3896 -0.1045
## 1020 54.3224 nan 0.3896 -0.1502
## 1040 53.8942 nan 0.3896 -0.0674
## 1060 53.4952 nan 0.3896 -0.1027
## 1080 53.0853 nan 0.3896 -0.1352
## 1100 52.5995 nan 0.3896 -0.1174
## 1120 52.1656 nan 0.3896 -0.1243
## 1140 51.7283 nan 0.3896 -0.0888
## 1160 51.3091 nan 0.3896 -0.1839
## 1180 50.8200 nan 0.3896 -0.1548
## 1200 50.4207 nan 0.3896 -0.1020
## 1220 50.0822 nan 0.3896 -0.0803
## 1240 49.7668 nan 0.3896 -0.1215
## 1260 49.3083 nan 0.3896 -0.0847
## 1280 49.0158 nan 0.3896 -0.0565
## 1300 48.6347 nan 0.3896 -0.0310
## 1320 48.1182 nan 0.3896 -0.1015
## 1340 47.8782 nan 0.3896 -0.1341
## 1360 47.6320 nan 0.3896 -0.2188
## 1380 47.3462 nan 0.3896 -0.0330
## 1400 47.0475 nan 0.3896 -0.2695
## 1420 46.7102 nan 0.3896 -0.0877
## 1440 46.2081 nan 0.3896 -0.0696
## 1460 45.8061 nan 0.3896 -0.0717
## 1480 45.4999 nan 0.3896 -0.0923
## 1500 45.2075 nan 0.3896 -0.1565
## 1520 44.9571 nan 0.3896 -0.0849
## 1540 44.8335 nan 0.3896 -0.1408
## 1560 44.5672 nan 0.3896 -0.2069
## 1580 44.3287 nan 0.3896 -0.0690
## 1600 44.0894 nan 0.3896 -0.1273
## 1620 43.8693 nan 0.3896 -0.1414
## 1640 43.6789 nan 0.3896 -0.1205
## 1660 43.3803 nan 0.3896 -0.1336
## 1680 43.1883 nan 0.3896 -0.0434
## 1700 43.0196 nan 0.3896 -0.0877
## 1720 42.6774 nan 0.3896 -0.0300
## 1740 42.5446 nan 0.3896 -0.1395
## 1760 42.3713 nan 0.3896 -0.0313
## 1780 42.1830 nan 0.3896 -0.0661
## 1800 41.9622 nan 0.3896 -0.0649
## 1820 41.8176 nan 0.3896 -0.0610
## 1840 41.6117 nan 0.3896 -0.0899
## 1860 41.3670 nan 0.3896 -0.1270
## 1880 41.2375 nan 0.3896 -0.2314
## 1900 41.0460 nan 0.3896 -0.0833
## 1920 40.8820 nan 0.3896 -0.1486
## 1940 40.6821 nan 0.3896 -0.0450
## 1960 40.4917 nan 0.3896 -0.1026
## 1980 40.2800 nan 0.3896 -0.0598
## 2000 40.0890 nan 0.3896 -0.0587
## 2020 39.8755 nan 0.3896 -0.1115
## 2040 39.7375 nan 0.3896 -0.1021
## 2060 39.6012 nan 0.3896 -0.0907
## 2080 39.5022 nan 0.3896 -0.0748
## 2100 39.3463 nan 0.3896 -0.0752
## 2120 39.2416 nan 0.3896 -0.1464
## 2140 39.0610 nan 0.3896 -0.1524
## 2160 38.8502 nan 0.3896 -0.0368
## 2180 38.6602 nan 0.3896 -0.0669
## 2200 38.4324 nan 0.3896 -0.0661
## 2220 38.2156 nan 0.3896 -0.0745
## 2240 38.0640 nan 0.3896 -0.1012
## 2260 37.9309 nan 0.3896 -0.0992
## 2280 37.7930 nan 0.3896 -0.0424
## 2300 37.6914 nan 0.3896 -0.0763
## 2320 37.4886 nan 0.3896 -0.0311
## 2340 37.3038 nan 0.3896 -0.0963
## 2360 37.1183 nan 0.3896 -0.0865
## 2380 37.0517 nan 0.3896 -0.1104
## 2400 36.9121 nan 0.3896 -0.0622
## 2420 36.8079 nan 0.3896 -0.0537
## 2440 36.6231 nan 0.3896 -0.0782
## 2460 36.4200 nan 0.3896 -0.0199
## 2480 36.3025 nan 0.3896 0.0098
## 2500 36.1643 nan 0.3896 -0.0845
## 2520 36.0804 nan 0.3896 -0.1302
## 2540 35.9442 nan 0.3896 -0.0949
## 2560 35.8479 nan 0.3896 -0.0959
## 2580 35.7413 nan 0.3896 -0.1038
## 2600 35.5764 nan 0.3896 -0.0774
## 2620 35.4393 nan 0.3896 -0.0918
## 2640 35.3948 nan 0.3896 -0.1244
## 2660 35.3002 nan 0.3896 -0.0834
## 2680 35.1201 nan 0.3896 -0.1994
## 2700 34.9479 nan 0.3896 -0.0880
## 2720 34.8885 nan 0.3896 -0.0662
## 2740 34.7747 nan 0.3896 -0.1117
## 2760 34.6756 nan 0.3896 -0.1346
## 2780 34.6000 nan 0.3896 -0.0417
## 2800 34.6503 nan 0.3896 -0.1389
## 2820 34.4205 nan 0.3896 -0.1587
## 2840 34.3233 nan 0.3896 -0.0770
## 2860 34.1981 nan 0.3896 -0.1535
## 2880 34.0937 nan 0.3896 -0.0960
## 2900 33.9943 nan 0.3896 -0.0507
## 2920 33.9042 nan 0.3896 -0.0412
## 2940 33.8041 nan 0.3896 -0.0979
## 2960 33.7884 nan 0.3896 -0.1181
## 2980 33.6237 nan 0.3896 -0.0506
## 3000 33.4821 nan 0.3896 -0.0732
## 3020 33.4736 nan 0.3896 -0.1440
## 3040 33.4127 nan 0.3896 -0.0753
## 3060 33.3178 nan 0.3896 -0.0322
## 3080 33.2760 nan 0.3896 -0.2064
## 3100 33.1716 nan 0.3896 -0.0657
## 3120 33.1371 nan 0.3896 -0.1217
## 3140 32.9607 nan 0.3896 -0.0618
## 3160 32.8767 nan 0.3896 -0.2242
## 3180 32.8222 nan 0.3896 -0.1226
## 3200 32.8001 nan 0.3896 -0.1425
## 3220 32.6850 nan 0.3896 -0.1300
## 3240 32.6208 nan 0.3896 -0.0329
## 3260 32.5128 nan 0.3896 -0.0512
## 3280 32.4430 nan 0.3896 -0.1764
## 3300 32.3674 nan 0.3896 -0.1104
## 3320 32.3272 nan 0.3896 -0.0456
## 3340 32.2889 nan 0.3896 -0.0825
## 3360 32.1970 nan 0.3896 -0.0818
## 3380 32.1566 nan 0.3896 -0.0593
## 3400 32.1000 nan 0.3896 -0.0829
## 3420 32.0453 nan 0.3896 -0.1719
## 3440 31.9537 nan 0.3896 -0.0599
## 3460 31.9128 nan 0.3896 -0.2856
## 3480 31.8516 nan 0.3896 -0.0653
## 3500 31.7987 nan 0.3896 -0.1114
## 3520 31.6687 nan 0.3896 -0.0457
## 3540 31.6340 nan 0.3896 -0.0677
## 3560 31.5510 nan 0.3896 -0.0599
## 3580 31.4923 nan 0.3896 -0.0924
## 3600 31.4246 nan 0.3896 -0.0682
## 3620 31.3499 nan 0.3896 -0.1036
## 3640 31.2919 nan 0.3896 -0.1012
## 3660 31.2134 nan 0.3896 -0.0649
## 3680 31.1513 nan 0.3896 -0.1660
## 3700 31.1048 nan 0.3896 -0.2067
## 3720 31.0415 nan 0.3896 -0.1493
## 3740 31.0269 nan 0.3896 -0.2850
## 3760 30.9477 nan 0.3896 -0.0107
## 3780 30.8868 nan 0.3896 -0.0735
## 3800 30.7940 nan 0.3896 -0.0767
## 3820 30.7705 nan 0.3896 -0.0397
## 3840 30.7294 nan 0.3896 -0.0413
## 3860 30.6902 nan 0.3896 -0.0963
## 3880 30.6040 nan 0.3896 -0.0383
## 3900 30.5430 nan 0.3896 -0.0772
## 3920 30.5378 nan 0.3896 -0.1499
## 3940 30.4630 nan 0.3896 -0.0624
## 3960 30.4285 nan 0.3896 -0.0625
## 3980 30.4085 nan 0.3896 -0.1534
## 4000 30.3510 nan 0.3896 -0.0691
## 4020 30.2927 nan 0.3896 -0.0372
## 4040 30.2426 nan 0.3896 -0.0861
## 4060 30.1657 nan 0.3896 -0.0670
## 4080 30.1677 nan 0.3896 -0.0924
##
## Iter TrainDeviance ValidDeviance StepSize Improve
## 1 542.3070 nan 0.5470 644.4317
## 2 368.4464 nan 0.5470 169.7810
## 3 314.3265 nan 0.5470 54.6140
## 4 290.2978 nan 0.5470 22.4680
## 5 271.5018 nan 0.5470 17.9091
## 6 259.1657 nan 0.5470 10.8291
## 7 250.7262 nan 0.5470 8.0476
## 8 240.5897 nan 0.5470 8.5867
## 9 232.6700 nan 0.5470 7.2057
## 10 229.2670 nan 0.5470 2.9190
## 20 193.5131 nan 0.5470 2.1525
## 40 161.0752 nan 0.5470 0.5130
## 60 143.5336 nan 0.5470 0.1071
## 80 131.8329 nan 0.5470 -0.2857
## 100 123.0206 nan 0.5470 -0.0716
## 120 116.1552 nan 0.5470 -0.0362
## 140 109.9345 nan 0.5470 0.1926
## 160 104.8267 nan 0.5470 0.1431
## 180 100.5709 nan 0.5470 -0.0387
## 200 97.4286 nan 0.5470 -0.3271
## 220 94.5115 nan 0.5470 -0.5509
## 240 91.2204 nan 0.5470 -0.0902
## 260 88.3644 nan 0.5470 -0.1467
## 280 85.8495 nan 0.5470 -0.1958
## 300 83.5324 nan 0.5470 -0.1665
## 320 81.4644 nan 0.5470 -0.3960
## 340 79.7604 nan 0.5470 -0.0978
## 360 78.1559 nan 0.5470 -0.1536
## 380 76.3635 nan 0.5470 -0.3493
## 400 74.4501 nan 0.5470 -0.2796
## 420 72.8697 nan 0.5470 -0.3021
## 440 71.3206 nan 0.5470 -0.2629
## 460 70.0988 nan 0.5470 -0.1268
## 480 68.7720 nan 0.5470 -0.2221
## 500 67.5797 nan 0.5470 -0.0668
## 520 66.5229 nan 0.5470 -0.1994
## 540 65.4074 nan 0.5470 -0.1591
## 560 64.3825 nan 0.5470 -0.2525
## 580 63.1884 nan 0.5470 -0.1506
## 600 62.1842 nan 0.5470 -0.1038
## 620 61.2088 nan 0.5470 0.0185
## 640 60.4141 nan 0.5470 -0.0729
## 660 59.6112 nan 0.5470 -0.1775
## 680 58.8692 nan 0.5470 -0.3279
## 700 58.0852 nan 0.5470 -0.1672
## 720 57.3576 nan 0.5470 -0.1369
## 740 56.5296 nan 0.5470 -0.1059
## 760 55.9631 nan 0.5470 -0.1367
## 780 55.4404 nan 0.5470 -0.2815
## 800 54.8895 nan 0.5470 -0.1879
## 820 54.0018 nan 0.5470 -0.0501
## 840 53.6397 nan 0.5470 -0.1549
## 860 53.0866 nan 0.5470 -0.1204
## 880 52.6250 nan 0.5470 -0.3210
## 900 51.9648 nan 0.5470 -0.1724
## 920 51.4949 nan 0.5470 -0.1707
## 940 50.9150 nan 0.5470 -0.1991
## 960 50.6005 nan 0.5470 -0.1655
## 980 50.1371 nan 0.5470 -0.1115
## 1000 49.6094 nan 0.5470 -0.1691
## 1020 49.1445 nan 0.5470 -0.0529
## 1040 48.7974 nan 0.5470 -0.0589
## 1060 48.4254 nan 0.5470 -0.2869
## 1080 48.1189 nan 0.5470 -0.2441
## 1100 47.6865 nan 0.5470 -0.1308
## 1120 47.2913 nan 0.5470 -0.1349
## 1140 46.8013 nan 0.5470 -0.1766
## 1160 46.7548 nan 0.5470 -0.0820
## 1180 46.2962 nan 0.5470 -0.1355
## 1200 46.1501 nan 0.5470 -0.1416
## 1220 45.9506 nan 0.5470 -0.1189
## 1240 45.5476 nan 0.5470 -0.2545
## 1260 45.1888 nan 0.5470 -0.1986
## 1280 44.9738 nan 0.5470 -0.2659
## 1300 44.5453 nan 0.5470 0.0091
## 1320 44.1553 nan 0.5470 -0.0806
## 1340 43.8385 nan 0.5470 -0.0776
## 1360 43.5605 nan 0.5470 -0.0254
## 1380 43.3799 nan 0.5470 -0.4751
## 1400 43.0405 nan 0.5470 -0.0629
## 1420 42.7333 nan 0.5470 -0.1647
## 1440 42.3977 nan 0.5470 -0.0333
## 1460 42.2396 nan 0.5470 -0.1350
## 1480 41.9200 nan 0.5470 -0.3009
## 1500 41.6715 nan 0.5470 -0.0658
## 1520 41.5951 nan 0.5470 -0.2379
## 1540 41.4340 nan 0.5470 -0.1704
## 1560 41.2784 nan 0.5470 -0.1565
## 1580 41.0612 nan 0.5470 -0.2533
## 1600 40.8879 nan 0.5470 -0.1364
## 1620 40.6279 nan 0.5470 -0.2405
## 1640 40.4515 nan 0.5470 -0.0808
## 1660 40.1216 nan 0.5470 -0.0216
## 1680 39.9889 nan 0.5470 -0.1354
## 1700 39.9035 nan 0.5470 -0.1826
## 1720 39.5808 nan 0.5470 -0.0906
## 1740 39.5627 nan 0.5470 -0.1038
## 1760 39.4605 nan 0.5470 -0.0630
## 1780 39.3401 nan 0.5470 -0.3842
## 1800 39.2815 nan 0.5470 -0.2805
## 1820 38.9755 nan 0.5470 -0.1910
## 1840 38.8255 nan 0.5470 -0.1247
## 1860 38.6672 nan 0.5470 -0.0948
## 1880 38.3884 nan 0.5470 -0.0536
## 1900 38.2184 nan 0.5470 -0.1244
## 1920 38.1689 nan 0.5470 -0.2206
## 1940 38.0343 nan 0.5470 -0.0785
## 1960 37.9118 nan 0.5470 -0.1620
## 1980 37.6805 nan 0.5470 -0.1042
## 2000 37.5742 nan 0.5470 -0.1149
## 2020 37.3648 nan 0.5470 -0.0607
## 2040 37.1868 nan 0.5470 -0.1555
## 2060 36.9664 nan 0.5470 -0.0938
## 2080 36.7324 nan 0.5470 -0.0859
## 2100 36.6836 nan 0.5470 -0.1191
## 2120 36.5359 nan 0.5470 -0.1499
## 2140 36.3803 nan 0.5470 -0.0909
## 2160 36.1621 nan 0.5470 -0.0565
## 2180 36.0219 nan 0.5470 -0.1781
## 2200 35.9661 nan 0.5470 -0.5330
## 2220 35.7435 nan 0.5470 -0.2108
## 2240 35.6967 nan 0.5470 -0.1849
## 2260 35.6375 nan 0.5470 -0.1349
## 2280 35.5614 nan 0.5470 -0.0648
## 2300 35.3967 nan 0.5470 -0.1883
## 2320 35.2976 nan 0.5470 -0.2238
## 2340 35.2129 nan 0.5470 -0.0572
## 2360 35.0857 nan 0.5470 -0.0937
## 2380 34.9542 nan 0.5470 -0.1301
## 2400 34.8235 nan 0.5470 -0.1073
## 2420 34.7509 nan 0.5470 -0.0557
## 2440 34.6844 nan 0.5470 -0.0720
## 2460 34.6298 nan 0.5470 -0.2208
## 2480 34.5822 nan 0.5470 -0.2594
## 2500 34.4290 nan 0.5470 -0.1676
## 2520 34.3626 nan 0.5470 -0.0657
## 2540 34.2781 nan 0.5470 -0.1690
## 2560 34.2299 nan 0.5470 -0.0764
## 2580 34.1203 nan 0.5470 -0.0327
## 2600 33.9658 nan 0.5470 -0.0627
## 2620 33.8997 nan 0.5470 -0.0979
## 2640 33.8462 nan 0.5470 -0.0764
## 2660 33.7370 nan 0.5470 -0.0954
## 2680 33.5634 nan 0.5470 -0.1037
## 2700 33.4722 nan 0.5470 -0.0976
## 2720 33.4428 nan 0.5470 -0.0932
## 2740 33.2766 nan 0.5470 -0.0669
## 2760 33.2764 nan 0.5470 -0.1596
## 2780 33.1801 nan 0.5470 -0.1805
## 2800 33.1706 nan 0.5470 -0.1152
## 2820 33.0487 nan 0.5470 -0.1551
## 2840 32.9630 nan 0.5470 -0.1115
## 2860 32.8965 nan 0.5470 -0.1768
## 2880 32.7566 nan 0.5470 -0.0684
## 2900 32.8322 nan 0.5470 -0.1971
## 2920 32.7676 nan 0.5470 -0.1966
## 2940 32.7351 nan 0.5470 -0.0640
## 2960 32.6732 nan 0.5470 -0.1380
## 2980 32.5712 nan 0.5470 -0.1326
## 3000 32.4169 nan 0.5470 -0.0455
## 3020 32.3694 nan 0.5470 -0.1154
## 3040 32.4450 nan 0.5470 -0.2046
## 3060 32.3691 nan 0.5470 -0.0721
## 3080 32.2814 nan 0.5470 -0.0554
## 3100 32.2736 nan 0.5470 -0.1928
## 3120 32.1975 nan 0.5470 -0.2064
## 3140 32.1136 nan 0.5470 -0.1993
## 3160 32.0175 nan 0.5470 -0.1553
## 3180 31.9881 nan 0.5470 -0.1636
## 3200 31.8310 nan 0.5470 -0.1240
## 3220 31.8190 nan 0.5470 -0.0703
## 3240 31.7363 nan 0.5470 -0.0993
## 3260 31.7272 nan 0.5470 -0.0916
## 3280 31.8131 nan 0.5470 -0.2832
## 3300 31.6560 nan 0.5470 -0.1187
## 3320 31.6394 nan 0.5470 -0.1642
## 3340 31.6659 nan 0.5470 -0.0913
## 3360 31.6340 nan 0.5470 -0.2551
## 3380 31.4976 nan 0.5470 -0.0999
## 3400 31.3762 nan 0.5470 -0.0962
## 3420 31.3078 nan 0.5470 -0.1735
## 3440 31.1764 nan 0.5470 -0.0921
## 3460 31.1662 nan 0.5470 -0.2471
## 3480 31.1405 nan 0.5470 -0.0582
## 3489 31.0313 nan 0.5470 -0.0522
##
## Iter TrainDeviance ValidDeviance StepSize Improve
## 1 821.3276 nan 0.2306 359.3408
## 2 600.9738 nan 0.2306 219.8959
## 3 460.3253 nan 0.2306 138.7767
## 4 372.6743 nan 0.2306 86.5394
## 5 318.0866 nan 0.2306 53.6244
## 6 280.0616 nan 0.2306 36.9895
## 7 251.1584 nan 0.2306 27.7687
## 8 230.8238 nan 0.2306 18.7941
## 9 216.7866 nan 0.2306 13.1819
## 10 207.0602 nan 0.2306 8.3991
## 20 163.1324 nan 0.2306 2.1658
## 40 131.7700 nan 0.2306 0.5595
## 60 115.2830 nan 0.2306 -0.0900
## 80 102.6829 nan 0.2306 0.2112
## 100 94.5790 nan 0.2306 0.0239
## 120 87.5823 nan 0.2306 -0.0845
## 140 82.2477 nan 0.2306 -0.2152
## 160 77.8893 nan 0.2306 -0.0167
## 180 74.1453 nan 0.2306 -0.2456
## 200 70.5234 nan 0.2306 -0.0206
## 220 67.4654 nan 0.2306 -0.1024
## 240 64.6399 nan 0.2306 -0.1322
## 260 62.1822 nan 0.2306 -0.1345
## 280 60.0635 nan 0.2306 -0.2072
## 300 57.9217 nan 0.2306 -0.1804
## 320 56.0681 nan 0.2306 -0.1069
## 340 54.4497 nan 0.2306 -0.1241
## 360 52.8719 nan 0.2306 -0.1624
## 380 51.5280 nan 0.2306 -0.1925
## 400 50.2007 nan 0.2306 -0.1414
## 420 49.0710 nan 0.2306 -0.1109
## 440 47.8819 nan 0.2306 -0.1300
## 460 46.7860 nan 0.2306 0.0172
## 480 45.6933 nan 0.2306 -0.1042
## 500 44.7527 nan 0.2306 -0.0893
## 520 43.8969 nan 0.2306 -0.1233
## 540 43.0154 nan 0.2306 -0.0938
## 560 42.1757 nan 0.2306 -0.1225
## 580 41.5056 nan 0.2306 -0.1356
## 600 40.9089 nan 0.2306 -0.0900
## 620 40.0992 nan 0.2306 -0.0755
## 640 39.4106 nan 0.2306 -0.1481
## 660 38.7655 nan 0.2306 -0.0154
## 680 38.1465 nan 0.2306 -0.1514
## 700 37.6731 nan 0.2306 -0.2110
## 720 37.1437 nan 0.2306 -0.0823
## 740 36.6343 nan 0.2306 -0.0714
## 760 36.2013 nan 0.2306 -0.1316
## 780 35.8551 nan 0.2306 -0.2566
## 800 35.4992 nan 0.2306 -0.1493
## 820 35.1056 nan 0.2306 -0.1832
## 840 34.6721 nan 0.2306 -0.1240
## 860 34.3447 nan 0.2306 -0.0989
## 880 33.9247 nan 0.2306 -0.0895
## 900 33.5514 nan 0.2306 -0.1365
## 920 33.2885 nan 0.2306 -0.1165
## 940 33.1227 nan 0.2306 -0.1423
## 960 32.8403 nan 0.2306 -0.0933
## 980 32.5868 nan 0.2306 -0.0878
## 1000 32.3950 nan 0.2306 -0.1110
## 1020 32.1584 nan 0.2306 -0.0874
## 1040 31.8960 nan 0.2306 -0.0693
## 1060 31.6905 nan 0.2306 -0.0807
## 1080 31.4874 nan 0.2306 -0.0493
## 1100 31.2477 nan 0.2306 -0.1203
## 1120 31.0578 nan 0.2306 -0.1281
## 1140 30.9094 nan 0.2306 -0.1814
## 1160 30.6055 nan 0.2306 -0.0216
## 1180 30.4484 nan 0.2306 -0.1130
## 1200 30.3480 nan 0.2306 -0.1480
## 1220 30.0757 nan 0.2306 -0.1390
## 1240 29.9736 nan 0.2306 -0.0968
## 1260 29.8186 nan 0.2306 -0.1288
## 1280 29.6686 nan 0.2306 -0.1193
## 1300 29.5137 nan 0.2306 -0.1118
## 1320 29.4115 nan 0.2306 -0.1184
## 1340 29.2279 nan 0.2306 -0.1556
## 1360 29.0795 nan 0.2306 -0.1155
## 1380 28.9351 nan 0.2306 -0.0774
## 1400 28.7856 nan 0.2306 -0.1238
## 1420 28.6727 nan 0.2306 -0.1252
## 1440 28.5775 nan 0.2306 -0.1711
## 1460 28.4776 nan 0.2306 -0.1192
## 1480 28.4216 nan 0.2306 -0.2796
## 1500 28.3189 nan 0.2306 -0.1458
## 1520 28.2425 nan 0.2306 -0.3081
## 1540 28.1180 nan 0.2306 -0.3040
## 1560 27.9131 nan 0.2306 -0.2382
## 1580 27.8117 nan 0.2306 -0.0584
## 1600 27.7041 nan 0.2306 -0.0846
## 1620 27.5870 nan 0.2306 -0.1009
## 1640 27.4971 nan 0.2306 -0.1263
## 1660 27.3889 nan 0.2306 -0.0867
## 1680 27.2859 nan 0.2306 -0.1151
## 1700 27.1732 nan 0.2306 -0.0785
## 1720 27.1091 nan 0.2306 -0.1300
## 1740 27.0780 nan 0.2306 -0.0984
## 1760 26.9646 nan 0.2306 -0.0605
## 1780 26.8318 nan 0.2306 -0.1066
## 1800 26.7576 nan 0.2306 -0.1019
## 1820 26.6844 nan 0.2306 -0.1053
## 1840 26.6453 nan 0.2306 -0.1630
## 1860 26.5627 nan 0.2306 -0.0971
## 1880 26.4985 nan 0.2306 -0.2263
## 1900 26.4268 nan 0.2306 -0.0818
## 1920 26.3996 nan 0.2306 -0.1320
## 1940 26.3506 nan 0.2306 -0.0988
## 1960 26.3046 nan 0.2306 -0.0680
## 1980 26.1787 nan 0.2306 -0.1133
## 2000 26.0859 nan 0.2306 -0.1469
## 2020 26.1298 nan 0.2306 -0.1239
## 2040 26.0477 nan 0.2306 -0.1247
## 2060 25.9427 nan 0.2306 -0.0734
## 2080 25.9105 nan 0.2306 -0.1723
## 2100 25.8732 nan 0.2306 -0.1463
## 2120 25.8305 nan 0.2306 -0.2124
## 2140 25.7787 nan 0.2306 -0.1266
## 2160 25.6507 nan 0.2306 -0.1174
## 2180 25.6593 nan 0.2306 -0.2485
## 2200 25.5332 nan 0.2306 -0.1312
## 2220 25.4667 nan 0.2306 -0.0913
## 2240 25.4175 nan 0.2306 -0.1287
## 2260 25.3485 nan 0.2306 -0.1188
## 2280 25.2882 nan 0.2306 -0.1097
## 2300 25.2328 nan 0.2306 -0.1159
## 2320 25.1848 nan 0.2306 -0.1273
## 2340 25.1040 nan 0.2306 -0.1248
## 2360 25.0788 nan 0.2306 -0.0993
## 2380 24.9987 nan 0.2306 -0.0883
## 2400 24.9645 nan 0.2306 -0.1105
## 2420 24.9213 nan 0.2306 -0.1001
## 2440 24.8627 nan 0.2306 -0.1023
## 2460 24.8351 nan 0.2306 -0.1412
## 2480 24.8188 nan 0.2306 -0.1868
## 2500 24.7693 nan 0.2306 -0.1497
## 2520 24.7641 nan 0.2306 -0.1357
## 2540 24.6546 nan 0.2306 -0.0896
## 2560 24.6410 nan 0.2306 -0.0583
## 2580 24.6273 nan 0.2306 -0.1648
## 2600 24.5793 nan 0.2306 -0.2057
## 2620 24.5329 nan 0.2306 -0.1209
## 2640 24.4869 nan 0.2306 -0.1038
## 2660 24.4552 nan 0.2306 -0.1387
## 2680 24.4255 nan 0.2306 -0.1795
## 2700 24.3308 nan 0.2306 -0.0570
## 2720 24.2750 nan 0.2306 -0.1012
## 2740 24.2412 nan 0.2306 -0.0967
## 2760 24.1883 nan 0.2306 -0.0974
## 2780 24.1515 nan 0.2306 -0.1412
## 2800 24.0884 nan 0.2306 -0.0207
## 2820 24.0719 nan 0.2306 -0.1365
## 2840 24.0523 nan 0.2306 -0.0712
## 2860 24.0244 nan 0.2306 -0.1552
## 2880 23.9892 nan 0.2306 -0.1515
## 2900 23.9778 nan 0.2306 -0.2321
## 2920 23.9419 nan 0.2306 -0.1755
## 2940 23.8970 nan 0.2306 -0.0948
## 2960 23.8516 nan 0.2306 -0.1394
## 2980 23.8075 nan 0.2306 -0.1650
## 3000 23.7923 nan 0.2306 -0.1693
## 3020 23.7382 nan 0.2306 -0.1186
## 3040 23.7196 nan 0.2306 -0.0865
## 3060 23.6355 nan 0.2306 -0.0840
## 3080 23.6224 nan 0.2306 -0.1374
## 3100 23.5921 nan 0.2306 -0.1078
## 3120 23.5343 nan 0.2306 -0.0601
## 3140 23.5309 nan 0.2306 -0.1842
## 3160 23.5316 nan 0.2306 -0.1278
## 3180 23.4955 nan 0.2306 -0.1127
## 3200 23.4582 nan 0.2306 -0.2115
## 3220 23.4536 nan 0.2306 -0.0731
## 3240 23.4456 nan 0.2306 -0.1514
## 3260 23.4634 nan 0.2306 -0.1580
## 3280 23.4093 nan 0.2306 -0.1692
## 3300 23.4322 nan 0.2306 -0.1611
## 3320 23.3961 nan 0.2306 -0.1979
## 3340 23.3659 nan 0.2306 -0.1000
## 3360 23.2976 nan 0.2306 -0.0943
## 3380 23.2648 nan 0.2306 -0.1334
## 3400 23.3039 nan 0.2306 -0.0970
## 3420 23.1956 nan 0.2306 -0.1461
## 3440 23.1778 nan 0.2306 -0.1615
## 3460 23.1878 nan 0.2306 -0.1024
## 3480 23.1336 nan 0.2306 -0.0868
## 3500 23.1256 nan 0.2306 -0.1345
## 3520 23.1319 nan 0.2306 -0.1503
## 3540 23.1050 nan 0.2306 -0.2077
## 3560 23.0862 nan 0.2306 -0.1774
## 3580 23.0154 nan 0.2306 -0.0965
## 3600 23.0157 nan 0.2306 -0.1314
## 3620 23.0188 nan 0.2306 -0.1554
## 3640 22.9422 nan 0.2306 -0.1125
## 3660 22.9363 nan 0.2306 -0.1066
## 3680 22.9579 nan 0.2306 -0.1537
## 3700 22.9607 nan 0.2306 -0.1180
## 3720 22.9389 nan 0.2306 -0.1488
##
## Iter TrainDeviance ValidDeviance StepSize Improve
## 1 676.8287 nan 0.3896 505.9714
## 2 467.9709 nan 0.3896 200.9572
## 3 373.1212 nan 0.3896 94.0872
## 4 322.0677 nan 0.3896 50.2540
## 5 298.3192 nan 0.3896 22.7675
## 6 278.9564 nan 0.3896 18.5713
## 7 267.4553 nan 0.3896 11.7286
## 8 258.4293 nan 0.3896 7.8794
## 9 249.7020 nan 0.3896 8.7866
## 10 243.4039 nan 0.3896 5.8770
## 20 204.9556 nan 0.3896 1.6074
## 40 171.7300 nan 0.3896 0.3857
## 60 154.7729 nan 0.3896 0.0859
## 80 141.1832 nan 0.3896 0.2308
## 100 131.9692 nan 0.3896 0.1222
## 120 124.8771 nan 0.3896 0.3768
## 140 119.6264 nan 0.3896 0.2125
## 160 114.5102 nan 0.3896 0.0814
## 180 110.3484 nan 0.3896 -0.1261
## 200 106.4667 nan 0.3896 0.0865
## 220 102.8844 nan 0.3896 -0.2331
## 240 99.7773 nan 0.3896 -0.1499
## 260 96.9362 nan 0.3896 0.0191
## 280 94.4762 nan 0.3896 -0.1258
## 300 92.6440 nan 0.3896 -0.2624
## 320 90.6197 nan 0.3896 -0.1723
## 340 88.8290 nan 0.3896 -0.1798
## 360 87.3736 nan 0.3896 -0.1580
## 380 85.4933 nan 0.3896 -0.0865
## 400 83.8877 nan 0.3896 -0.0746
## 420 82.1145 nan 0.3896 -0.0495
## 440 80.7048 nan 0.3896 0.0332
## 460 79.4790 nan 0.3896 -0.1341
## 480 78.5778 nan 0.3896 -0.1757
## 500 77.1777 nan 0.3896 -0.0136
## 520 76.0413 nan 0.3896 -0.0630
## 540 75.0136 nan 0.3896 -0.2771
## 560 73.9220 nan 0.3896 -0.1799
## 580 72.7934 nan 0.3896 -0.1094
## 600 71.9176 nan 0.3896 -0.0624
## 620 70.5897 nan 0.3896 -0.0914
## 640 69.6187 nan 0.3896 -0.1436
## 660 68.8592 nan 0.3896 -0.0062
## 680 68.0330 nan 0.3896 -0.0406
## 700 67.0629 nan 0.3896 -0.2016
## 720 66.2304 nan 0.3896 -0.1173
## 740 65.5563 nan 0.3896 -0.0605
## 760 64.5928 nan 0.3896 -0.0854
## 780 64.0018 nan 0.3896 -0.1374
## 800 63.3399 nan 0.3896 -0.1406
## 820 62.4860 nan 0.3896 -0.0692
## 840 61.9891 nan 0.3896 -0.0629
## 860 61.1159 nan 0.3896 -0.0343
## 880 60.5646 nan 0.3896 -0.0758
## 900 60.0431 nan 0.3896 -0.1246
## 920 59.3240 nan 0.3896 -0.1377
## 940 58.8028 nan 0.3896 -0.1204
## 960 58.1571 nan 0.3896 -0.0764
## 980 57.6320 nan 0.3896 -0.1426
## 1000 57.2725 nan 0.3896 -0.1796
## 1020 56.7767 nan 0.3896 -0.2723
## 1040 56.3310 nan 0.3896 -0.1124
## 1060 55.9910 nan 0.3896 -0.2214
## 1080 55.4359 nan 0.3896 -0.2438
## 1100 54.9728 nan 0.3896 -0.1382
## 1120 54.5233 nan 0.3896 -0.1839
## 1140 54.0115 nan 0.3896 -0.1016
## 1160 53.6043 nan 0.3896 -0.0981
## 1180 53.2291 nan 0.3896 -0.1246
## 1200 52.9151 nan 0.3896 -0.0252
## 1220 52.5999 nan 0.3896 -0.0890
## 1240 52.1093 nan 0.3896 -0.1119
## 1260 51.8856 nan 0.3896 -0.2350
## 1280 51.5313 nan 0.3896 -0.1463
## 1300 51.2772 nan 0.3896 -0.1585
## 1320 50.9728 nan 0.3896 -0.1015
## 1340 50.7128 nan 0.3896 -0.1269
## 1360 50.4766 nan 0.3896 -0.0812
## 1380 50.2285 nan 0.3896 -0.2744
## 1400 49.9840 nan 0.3896 -0.1105
## 1420 49.7098 nan 0.3896 -0.1211
## 1440 49.3795 nan 0.3896 -0.0909
## 1460 49.1345 nan 0.3896 -0.1356
## 1480 48.9698 nan 0.3896 -0.1806
## 1500 48.6178 nan 0.3896 -0.1611
## 1520 48.3375 nan 0.3896 -0.1153
## 1540 48.0315 nan 0.3896 -0.0840
## 1560 47.7846 nan 0.3896 -0.0894
## 1580 47.5932 nan 0.3896 -0.1034
## 1600 47.3227 nan 0.3896 -0.0601
## 1620 47.1003 nan 0.3896 -0.1954
## 1640 46.8264 nan 0.3896 -0.1180
## 1660 46.6298 nan 0.3896 -0.0623
## 1680 46.3883 nan 0.3896 -0.1209
## 1700 46.2756 nan 0.3896 -0.0784
## 1720 46.0333 nan 0.3896 -0.1305
## 1740 45.9145 nan 0.3896 -0.1290
## 1760 45.7022 nan 0.3896 -0.0923
## 1780 45.4369 nan 0.3896 -0.0985
## 1800 45.2338 nan 0.3896 -0.0775
## 1820 44.9947 nan 0.3896 -0.0811
## 1840 44.8071 nan 0.3896 -0.0899
## 1860 44.5211 nan 0.3896 -0.1377
## 1880 44.3113 nan 0.3896 -0.0899
## 1900 44.1861 nan 0.3896 -0.1104
## 1920 43.9825 nan 0.3896 -0.1269
## 1940 43.7299 nan 0.3896 -0.0252
## 1960 43.5450 nan 0.3896 -0.1480
## 1980 43.2532 nan 0.3896 -0.1382
## 2000 43.1142 nan 0.3896 -0.1281
## 2020 42.8611 nan 0.3896 -0.0760
## 2040 42.7771 nan 0.3896 -0.2003
## 2060 42.6444 nan 0.3896 -0.0250
## 2080 42.5149 nan 0.3896 -0.1468
## 2100 42.3439 nan 0.3896 -0.0920
## 2120 42.1556 nan 0.3896 -0.2042
## 2140 42.0129 nan 0.3896 -0.0920
## 2160 41.8834 nan 0.3896 -0.0705
## 2180 41.7364 nan 0.3896 -0.1326
## 2200 41.5859 nan 0.3896 -0.0704
## 2220 41.4262 nan 0.3896 -0.0961
## 2240 41.3361 nan 0.3896 -0.1172
## 2260 41.2428 nan 0.3896 -0.1088
## 2280 41.3340 nan 0.3896 -0.3704
## 2300 41.0908 nan 0.3896 -0.0875
## 2320 40.9103 nan 0.3896 -0.0691
## 2340 40.7466 nan 0.3896 -0.1031
## 2360 40.6143 nan 0.3896 -0.0597
## 2380 40.4672 nan 0.3896 -0.1510
## 2400 40.3150 nan 0.3896 -0.0669
## 2420 40.1931 nan 0.3896 -0.1276
## 2440 40.1820 nan 0.3896 -0.1633
## 2460 40.0486 nan 0.3896 -0.2673
## 2480 39.9413 nan 0.3896 -0.1272
## 2500 39.8281 nan 0.3896 -0.1158
## 2520 39.6440 nan 0.3896 -0.1168
## 2540 39.5155 nan 0.3896 -0.0856
## 2560 39.4196 nan 0.3896 -0.0708
## 2580 39.3138 nan 0.3896 -0.0960
## 2600 39.1487 nan 0.3896 -0.0573
## 2620 39.0594 nan 0.3896 -0.0649
## 2640 38.9706 nan 0.3896 -0.3441
## 2660 38.7677 nan 0.3896 -0.0537
## 2680 38.7314 nan 0.3896 -0.0902
## 2700 38.5691 nan 0.3896 -0.0750
## 2720 38.4278 nan 0.3896 -0.0386
## 2740 38.3359 nan 0.3896 -0.1170
## 2760 38.2188 nan 0.3896 -0.1087
## 2780 38.1124 nan 0.3896 -0.0449
## 2800 37.9610 nan 0.3896 -0.0090
## 2820 37.8863 nan 0.3896 -0.0674
## 2840 37.7188 nan 0.3896 -0.0977
## 2860 37.6552 nan 0.3896 -0.1190
## 2880 37.5271 nan 0.3896 -0.1117
## 2900 37.4721 nan 0.3896 -0.1001
## 2920 37.3468 nan 0.3896 -0.0524
## 2940 37.2613 nan 0.3896 -0.0907
## 2960 37.1944 nan 0.3896 -0.1266
## 2980 37.0689 nan 0.3896 -0.1222
## 3000 37.0133 nan 0.3896 -0.0981
## 3020 36.9193 nan 0.3896 -0.1368
## 3040 36.8685 nan 0.3896 -0.1411
## 3060 36.7645 nan 0.3896 -0.0748
## 3080 36.7170 nan 0.3896 -0.0859
## 3100 36.7008 nan 0.3896 -0.0898
## 3120 36.5408 nan 0.3896 -0.1598
## 3140 36.4719 nan 0.3896 -0.0814
## 3160 36.3800 nan 0.3896 -0.0698
## 3180 36.2939 nan 0.3896 -0.0961
## 3200 36.2154 nan 0.3896 -0.1578
## 3220 36.1782 nan 0.3896 -0.0864
## 3240 36.1416 nan 0.3896 -0.1789
## 3260 36.0564 nan 0.3896 -0.0554
## 3280 36.0000 nan 0.3896 -0.1661
## 3300 35.8914 nan 0.3896 -0.0586
## 3320 35.7453 nan 0.3896 -0.1078
## 3340 35.6428 nan 0.3896 -0.0919
## 3360 35.4969 nan 0.3896 -0.0816
## 3380 35.4574 nan 0.3896 -0.0926
## 3400 35.3927 nan 0.3896 -0.0857
## 3420 35.3836 nan 0.3896 -0.0902
## 3440 35.3137 nan 0.3896 -0.0589
## 3460 35.2626 nan 0.3896 -0.0031
## 3480 35.2074 nan 0.3896 -0.1331
## 3500 35.1717 nan 0.3896 -0.1036
## 3520 35.1323 nan 0.3896 -0.1364
## 3540 35.0813 nan 0.3896 -0.0705
## 3560 34.9507 nan 0.3896 -0.0687
## 3580 34.9035 nan 0.3896 -0.0818
## 3600 34.9124 nan 0.3896 -0.2005
## 3620 34.8530 nan 0.3896 -0.0495
## 3640 34.7578 nan 0.3896 -0.1039
## 3660 34.7142 nan 0.3896 -0.0516
## 3680 34.6825 nan 0.3896 -0.1496
## 3700 34.5440 nan 0.3896 -0.0473
## 3720 34.5356 nan 0.3896 -0.0817
## 3740 34.4491 nan 0.3896 -0.1598
## 3760 34.4218 nan 0.3896 -0.0919
## 3780 34.3640 nan 0.3896 -0.0673
## 3800 34.2794 nan 0.3896 -0.0990
## 3820 34.2235 nan 0.3896 -0.1252
## 3840 34.1632 nan 0.3896 -0.1456
## 3860 34.0918 nan 0.3896 -0.1276
## 3880 34.0270 nan 0.3896 -0.0561
## 3900 33.9516 nan 0.3896 -0.0563
## 3920 33.8951 nan 0.3896 -0.0384
## 3940 33.8385 nan 0.3896 -0.1006
## 3960 33.7992 nan 0.3896 -0.0723
## 3980 33.6824 nan 0.3896 -0.0624
## 4000 33.5855 nan 0.3896 -0.0838
## 4020 33.5784 nan 0.3896 -0.1084
## 4040 33.5179 nan 0.3896 -0.0573
## 4060 33.4657 nan 0.3896 -0.0838
## 4080 33.3659 nan 0.3896 -0.1464
##
## Iter TrainDeviance ValidDeviance StepSize Improve
## 1 543.2902 nan 0.5470 642.9670
## 2 373.2822 nan 0.5470 161.9944
## 3 313.2520 nan 0.5470 60.3204
## 4 289.9231 nan 0.5470 21.5888
## 5 272.8668 nan 0.5470 14.6496
## 6 259.6409 nan 0.5470 12.8673
## 7 250.7581 nan 0.5470 8.3044
## 8 242.4308 nan 0.5470 6.5270
## 9 235.7331 nan 0.5470 5.5650
## 10 229.3432 nan 0.5470 4.3093
## 20 195.5961 nan 0.5470 1.3428
## 40 163.9721 nan 0.5470 0.3904
## 60 146.2753 nan 0.5470 0.1279
## 80 133.3696 nan 0.5470 0.2029
## 100 125.5669 nan 0.5470 -0.1192
## 120 118.6052 nan 0.5470 -0.2355
## 140 112.7494 nan 0.5470 -0.3300
## 160 107.8631 nan 0.5470 -0.2354
## 180 104.1517 nan 0.5470 -0.6006
## 200 100.8986 nan 0.5470 -0.3900
## 220 97.5433 nan 0.5470 -0.5273
## 240 94.5163 nan 0.5470 -0.1667
## 260 92.1952 nan 0.5470 -0.1039
## 280 89.7928 nan 0.5470 -0.1748
## 300 87.5074 nan 0.5470 -0.3275
## 320 85.9551 nan 0.5470 -0.0629
## 340 83.7395 nan 0.5470 -0.1689
## 360 82.3494 nan 0.5470 -0.3428
## 380 80.6743 nan 0.5470 -0.2775
## 400 78.8831 nan 0.5470 -0.1820
## 420 77.3032 nan 0.5470 -0.0858
## 440 75.8846 nan 0.5470 -0.4080
## 460 74.7209 nan 0.5470 -0.1814
## 480 73.4075 nan 0.5470 -0.2105
## 500 72.2259 nan 0.5470 -0.1348
## 520 71.1456 nan 0.5470 -0.1489
## 540 70.1142 nan 0.5470 -0.1355
## 560 69.3827 nan 0.5470 -0.2710
## 580 68.4950 nan 0.5470 -0.2422
## 600 67.4828 nan 0.5470 -0.4141
## 620 66.8674 nan 0.5470 -0.2572
## 640 66.1788 nan 0.5470 -0.2728
## 660 65.2337 nan 0.5470 -0.3546
## 680 64.4266 nan 0.5470 -0.2667
## 700 63.7410 nan 0.5470 -0.1275
## 720 63.2553 nan 0.5470 -0.1793
## 740 62.4586 nan 0.5470 -0.2372
## 760 61.7030 nan 0.5470 -0.2507
## 780 61.0215 nan 0.5470 -0.2345
## 800 60.2494 nan 0.5470 -0.1153
## 820 59.7484 nan 0.5470 -0.2199
## 840 58.9687 nan 0.5470 -0.2443
## 860 58.6195 nan 0.5470 -0.1420
## 880 58.1017 nan 0.5470 -0.1680
## 900 57.5742 nan 0.5470 -0.2305
## 920 57.1509 nan 0.5470 -0.1647
## 940 56.6896 nan 0.5470 -0.2911
## 960 56.2129 nan 0.5470 -0.0576
## 980 55.9502 nan 0.5470 -0.2753
## 1000 55.5421 nan 0.5470 -0.0968
## 1020 55.0909 nan 0.5470 -0.2751
## 1040 54.6635 nan 0.5470 -0.2257
## 1060 54.3639 nan 0.5470 -0.0943
## 1080 53.9924 nan 0.5470 -0.1386
## 1100 53.6179 nan 0.5470 -0.1257
## 1120 53.1353 nan 0.5470 -0.0993
## 1140 52.7589 nan 0.5470 -0.3723
## 1160 52.4177 nan 0.5470 -0.0920
## 1180 52.1075 nan 0.5470 -0.4182
## 1200 51.7874 nan 0.5470 -0.1833
## 1220 51.3800 nan 0.5470 -0.3128
## 1240 50.8858 nan 0.5470 -0.0867
## 1260 50.6231 nan 0.5470 -0.1651
## 1280 50.4502 nan 0.5470 -0.2353
## 1300 50.1398 nan 0.5470 -0.1726
## 1320 49.8098 nan 0.5470 -0.0456
## 1340 49.5186 nan 0.5470 -0.2296
## 1360 49.2603 nan 0.5470 -0.1637
## 1380 48.8592 nan 0.5470 -0.2014
## 1400 48.5384 nan 0.5470 -0.1874
## 1420 48.2950 nan 0.5470 -0.2182
## 1440 47.9710 nan 0.5470 -0.1755
## 1460 47.8083 nan 0.5470 -0.2109
## 1480 47.6721 nan 0.5470 -0.1533
## 1500 47.2083 nan 0.5470 -0.0607
## 1520 47.0553 nan 0.5470 -0.1543
## 1540 46.8502 nan 0.5470 -0.1077
## 1560 46.6412 nan 0.5470 -0.0965
## 1580 46.5228 nan 0.5470 -0.1653
## 1600 46.2853 nan 0.5470 -0.1785
## 1620 46.0030 nan 0.5470 -0.0922
## 1640 45.7408 nan 0.5470 -0.0234
## 1660 45.4373 nan 0.5470 -0.1322
## 1680 45.3572 nan 0.5470 -0.3595
## 1700 45.0745 nan 0.5470 -0.0944
## 1720 44.8740 nan 0.5470 -0.0694
## 1740 44.8181 nan 0.5470 -0.1782
## 1760 44.6851 nan 0.5470 -0.1465
## 1780 44.4759 nan 0.5470 -0.1182
## 1800 44.3284 nan 0.5470 -0.1627
## 1820 44.2490 nan 0.5470 -0.1362
## 1840 44.0351 nan 0.5470 -0.1079
## 1860 43.8330 nan 0.5470 -0.1418
## 1880 43.5797 nan 0.5470 -0.1219
## 1900 43.4890 nan 0.5470 -0.1027
## 1920 43.3024 nan 0.5470 -0.1515
## 1940 43.1346 nan 0.5470 -0.1033
## 1960 43.0272 nan 0.5470 -0.1601
## 1980 42.8972 nan 0.5470 -0.0687
## 2000 42.7543 nan 0.5470 -0.0923
## 2020 42.5170 nan 0.5470 -0.0971
## 2040 42.3054 nan 0.5470 -0.0910
## 2060 42.0507 nan 0.5470 -0.1613
## 2080 41.9463 nan 0.5470 -0.1255
## 2100 41.8276 nan 0.5470 -0.2771
## 2120 41.6204 nan 0.5470 -0.1755
## 2140 41.5846 nan 0.5470 -0.2069
## 2160 41.3908 nan 0.5470 -0.2599
## 2180 41.1993 nan 0.5470 -0.1865
## 2200 40.8918 nan 0.5470 -0.0519
## 2220 40.8622 nan 0.5470 -0.1517
## 2240 40.6850 nan 0.5470 -0.0821
## 2260 40.5959 nan 0.5470 -0.3322
## 2280 40.2669 nan 0.5470 -0.0506
## 2300 40.2046 nan 0.5470 -0.1228
## 2320 40.0961 nan 0.5470 -0.2448
## 2340 40.0020 nan 0.5470 -0.2002
## 2360 40.0348 nan 0.5470 -0.1346
## 2380 40.0184 nan 0.5470 -0.2327
## 2400 39.9439 nan 0.5470 -0.1842
## 2420 39.7557 nan 0.5470 -0.1474
## 2440 39.5652 nan 0.5470 -0.1239
## 2460 39.4805 nan 0.5470 -0.2741
## 2480 39.5025 nan 0.5470 -0.2562
## 2500 39.3105 nan 0.5470 -0.1870
## 2520 39.1510 nan 0.5470 -0.1515
## 2540 39.0858 nan 0.5470 -0.1792
## 2560 38.9945 nan 0.5470 -0.2104
## 2580 38.6985 nan 0.5470 -0.1916
## 2600 38.6645 nan 0.5470 -0.2186
## 2620 38.4867 nan 0.5470 -0.0851
## 2640 38.3352 nan 0.5470 -0.1307
## 2660 38.2832 nan 0.5470 -0.0775
## 2680 38.1380 nan 0.5470 -0.2111
## 2700 38.0856 nan 0.5470 -0.1089
## 2720 38.0207 nan 0.5470 -0.3192
## 2740 37.9320 nan 0.5470 -0.0061
## 2760 37.8371 nan 0.5470 -0.1631
## 2780 37.7681 nan 0.5470 -0.1179
## 2800 37.7008 nan 0.5470 -0.1577
## 2820 37.6497 nan 0.5470 -0.2364
## 2840 37.5277 nan 0.5470 -0.1984
## 2860 37.4864 nan 0.5470 -0.2001
## 2880 37.3232 nan 0.5470 -0.1487
## 2900 37.1842 nan 0.5470 -0.0683
## 2920 37.1141 nan 0.5470 -0.1939
## 2940 37.0779 nan 0.5470 -0.1367
## 2960 37.1065 nan 0.5470 -0.3537
## 2980 37.0875 nan 0.5470 -0.2776
## 3000 36.9304 nan 0.5470 -0.0910
## 3020 36.8365 nan 0.5470 -0.1062
## 3040 36.6602 nan 0.5470 -0.1549
## 3060 36.6713 nan 0.5470 -0.2063
## 3080 36.5461 nan 0.5470 -0.1155
## 3100 36.5288 nan 0.5470 -0.1784
## 3120 36.4955 nan 0.5470 -0.1031
## 3140 36.3815 nan 0.5470 -0.1499
## 3160 36.3068 nan 0.5470 -0.1878
## 3180 36.4305 nan 0.5470 -0.1305
## 3200 36.2651 nan 0.5470 -0.1855
## 3220 36.2261 nan 0.5470 -0.2551
## 3240 36.2460 nan 0.5470 -0.2845
## 3260 36.1162 nan 0.5470 -0.1622
## 3280 36.0146 nan 0.5470 -0.1623
## 3300 35.9886 nan 0.5470 -0.1559
## 3320 35.9898 nan 0.5470 -0.2043
## 3340 35.8479 nan 0.5470 -0.1480
## 3360 35.8408 nan 0.5470 -0.0746
## 3380 35.6562 nan 0.5470 -0.0437
## 3400 35.5697 nan 0.5470 -0.1509
## 3420 35.5071 nan 0.5470 -0.1226
## 3440 35.4057 nan 0.5470 -0.2016
## 3460 35.3219 nan 0.5470 -0.1464
## 3480 35.2542 nan 0.5470 -0.1468
## 3489 35.2627 nan 0.5470 -0.2350
##
## Iter TrainDeviance ValidDeviance StepSize Improve
## 1 820.6194 nan 0.2306 353.8688
## 2 599.1583 nan 0.2306 217.2470
## 3 460.5915 nan 0.2306 138.5036
## 4 369.1621 nan 0.2306 88.4547
## 5 309.9012 nan 0.2306 56.9300
## 6 271.6850 nan 0.2306 37.6357
## 7 245.2320 nan 0.2306 25.4972
## 8 228.3411 nan 0.2306 15.3058
## 9 213.9316 nan 0.2306 13.3273
## 10 204.2950 nan 0.2306 8.5204
## 20 161.5396 nan 0.2306 1.7651
## 40 133.4623 nan 0.2306 0.7274
## 60 117.1551 nan 0.2306 -0.0256
## 80 105.1088 nan 0.2306 -0.1494
## 100 96.1503 nan 0.2306 -0.0269
## 120 89.3925 nan 0.2306 -0.0719
## 140 84.0294 nan 0.2306 -0.1394
## 160 79.7003 nan 0.2306 -0.0210
## 180 75.6240 nan 0.2306 -0.0973
## 200 71.5431 nan 0.2306 -0.0922
## 220 68.4637 nan 0.2306 -0.1434
## 240 65.3949 nan 0.2306 -0.1707
## 260 62.9546 nan 0.2306 -0.0553
## 280 60.6517 nan 0.2306 -0.0950
## 300 58.5007 nan 0.2306 -0.1367
## 320 56.6252 nan 0.2306 -0.1769
## 340 54.6718 nan 0.2306 -0.0940
## 360 52.9685 nan 0.2306 -0.0658
## 380 51.2805 nan 0.2306 -0.0203
## 400 50.1273 nan 0.2306 -0.0502
## 420 48.8870 nan 0.2306 -0.1049
## 440 47.7696 nan 0.2306 -0.1053
## 460 46.8493 nan 0.2306 -0.1287
## 480 45.7117 nan 0.2306 -0.1242
## 500 44.8263 nan 0.2306 -0.1519
## 520 43.8201 nan 0.2306 -0.1555
## 540 42.8759 nan 0.2306 -0.1274
## 560 42.1196 nan 0.2306 -0.1254
## 580 41.4654 nan 0.2306 -0.1650
## 600 40.8042 nan 0.2306 -0.2020
## 620 40.0755 nan 0.2306 -0.1853
## 640 39.4831 nan 0.2306 -0.1462
## 660 38.8501 nan 0.2306 -0.1288
## 680 38.3266 nan 0.2306 -0.1538
## 700 37.7959 nan 0.2306 -0.0576
## 720 37.2940 nan 0.2306 -0.1725
## 740 36.8893 nan 0.2306 -0.1297
## 760 36.3590 nan 0.2306 -0.1548
## 780 35.8777 nan 0.2306 -0.2197
## 800 35.5212 nan 0.2306 -0.1608
## 820 35.0515 nan 0.2306 -0.1663
## 840 34.6376 nan 0.2306 -0.1291
## 860 34.2711 nan 0.2306 -0.0653
## 880 33.9617 nan 0.2306 -0.0885
## 900 33.7386 nan 0.2306 -0.0681
## 920 33.3970 nan 0.2306 -0.1269
## 940 33.0884 nan 0.2306 -0.0888
## 960 32.8715 nan 0.2306 -0.0974
## 980 32.5560 nan 0.2306 -0.0833
## 1000 32.2511 nan 0.2306 -0.1256
## 1020 31.9856 nan 0.2306 -0.1033
## 1040 31.7123 nan 0.2306 -0.0611
## 1060 31.4901 nan 0.2306 -0.1139
## 1080 31.2426 nan 0.2306 -0.1382
## 1100 31.0234 nan 0.2306 -0.0596
## 1120 30.7606 nan 0.2306 -0.1187
## 1140 30.5245 nan 0.2306 -0.1529
## 1160 30.3948 nan 0.2306 -0.0725
## 1180 30.1445 nan 0.2306 -0.1676
## 1200 29.9077 nan 0.2306 -0.0736
## 1220 29.7746 nan 0.2306 -0.1501
## 1240 29.5787 nan 0.2306 -0.0625
## 1260 29.4524 nan 0.2306 -0.1050
## 1280 29.2022 nan 0.2306 -0.0740
## 1300 29.0184 nan 0.2306 -0.1378
## 1320 28.8193 nan 0.2306 -0.0367
## 1340 28.7192 nan 0.2306 -0.1775
## 1360 28.5234 nan 0.2306 -0.1312
## 1380 28.3592 nan 0.2306 -0.0806
## 1400 28.2234 nan 0.2306 -0.1832
## 1420 28.0449 nan 0.2306 -0.0678
## 1440 27.9516 nan 0.2306 -0.1192
## 1460 27.8408 nan 0.2306 -0.0580
## 1480 27.7067 nan 0.2306 -0.1908
## 1500 27.5488 nan 0.2306 -0.0931
## 1520 27.4258 nan 0.2306 -0.2191
## 1540 27.3084 nan 0.2306 -0.0845
## 1560 27.2389 nan 0.2306 -0.0935
## 1580 27.1365 nan 0.2306 -0.1586
## 1600 26.9924 nan 0.2306 -0.2679
## 1620 26.9321 nan 0.2306 -0.0944
## 1640 26.8169 nan 0.2306 -0.0948
## 1660 26.7395 nan 0.2306 -0.1229
## 1680 26.6600 nan 0.2306 -0.1464
## 1700 26.5451 nan 0.2306 -0.1251
## 1720 26.4734 nan 0.2306 -0.0913
## 1740 26.4018 nan 0.2306 -0.0413
## 1760 26.3240 nan 0.2306 -0.1701
## 1780 26.2329 nan 0.2306 -0.0909
## 1800 26.1518 nan 0.2306 -0.0868
## 1820 26.1005 nan 0.2306 -0.1193
## 1840 25.9644 nan 0.2306 -0.1858
## 1860 25.9329 nan 0.2306 -0.2324
## 1880 25.8601 nan 0.2306 -0.1522
## 1900 25.7788 nan 0.2306 -0.0746
## 1920 25.7110 nan 0.2306 -0.0837
## 1940 25.6263 nan 0.2306 -0.1669
## 1960 25.6022 nan 0.2306 -0.2031
## 1980 25.5498 nan 0.2306 -0.0747
## 2000 25.4881 nan 0.2306 -0.1205
## 2020 25.4198 nan 0.2306 -0.1501
## 2040 25.3321 nan 0.2306 -0.1091
## 2060 25.3483 nan 0.2306 -0.1233
## 2080 25.2751 nan 0.2306 -0.1638
## 2100 25.1988 nan 0.2306 -0.1019
## 2120 25.1271 nan 0.2306 -0.0669
## 2140 25.0705 nan 0.2306 -0.0881
## 2160 24.9968 nan 0.2306 -0.1178
## 2180 24.9626 nan 0.2306 -0.1475
## 2200 24.9035 nan 0.2306 -0.1245
## 2220 24.8270 nan 0.2306 -0.1577
## 2240 24.8116 nan 0.2306 -0.1813
## 2260 24.7574 nan 0.2306 -0.2410
## 2280 24.6536 nan 0.2306 -0.0455
## 2300 24.6072 nan 0.2306 -0.0927
## 2320 24.5196 nan 0.2306 -0.1021
## 2340 24.4471 nan 0.2306 -0.1662
## 2360 24.3710 nan 0.2306 -0.0817
## 2380 24.3157 nan 0.2306 -0.1340
## 2400 24.2798 nan 0.2306 -0.1065
## 2420 24.2078 nan 0.2306 -0.0971
## 2440 24.1877 nan 0.2306 -0.0973
## 2460 24.1814 nan 0.2306 -0.2632
## 2480 24.0947 nan 0.2306 -0.1100
## 2500 24.0599 nan 0.2306 -0.0619
## 2520 24.0449 nan 0.2306 -0.0418
## 2540 23.9912 nan 0.2306 -0.1012
## 2560 23.9605 nan 0.2306 -0.0675
## 2580 23.8798 nan 0.2306 -0.1198
## 2600 23.8217 nan 0.2306 -0.0890
## 2620 23.8019 nan 0.2306 -0.1085
## 2640 23.7905 nan 0.2306 -0.1211
## 2660 23.7463 nan 0.2306 -0.1934
## 2680 23.7228 nan 0.2306 -0.0635
## 2700 23.6900 nan 0.2306 -0.0781
## 2720 23.6591 nan 0.2306 -0.1247
## 2740 23.6356 nan 0.2306 -0.1673
## 2760 23.6248 nan 0.2306 -0.1023
## 2780 23.5742 nan 0.2306 -0.1595
## 2800 23.5516 nan 0.2306 -0.2022
## 2820 23.5146 nan 0.2306 -0.1024
## 2840 23.4404 nan 0.2306 -0.1018
## 2860 23.3906 nan 0.2306 -0.0990
## 2880 23.3689 nan 0.2306 -0.1056
## 2900 23.3536 nan 0.2306 -0.1049
## 2920 23.3225 nan 0.2306 -0.0685
## 2940 23.3121 nan 0.2306 -0.1098
## 2960 23.2444 nan 0.2306 -0.1059
## 2980 23.2215 nan 0.2306 -0.1258
## 3000 23.2125 nan 0.2306 -0.0950
## 3020 23.1764 nan 0.2306 -0.1731
## 3040 23.1885 nan 0.2306 -0.0953
## 3060 23.1645 nan 0.2306 -0.1524
## 3080 23.1262 nan 0.2306 -0.0670
## 3100 23.1659 nan 0.2306 -0.1602
## 3120 23.0794 nan 0.2306 -0.0874
## 3140 23.0536 nan 0.2306 -0.1239
## 3160 23.0068 nan 0.2306 -0.0622
## 3180 22.9841 nan 0.2306 -0.1110
## 3200 22.9964 nan 0.2306 -0.1033
## 3220 22.9546 nan 0.2306 -0.1292
## 3240 22.9308 nan 0.2306 -0.1357
## 3260 22.8933 nan 0.2306 -0.0912
## 3280 22.8573 nan 0.2306 -0.1191
## 3300 22.8943 nan 0.2306 -0.0955
## 3320 22.8265 nan 0.2306 -0.1671
## 3340 22.8042 nan 0.2306 -0.1781
## 3360 22.7992 nan 0.2306 -0.1349
## 3380 22.7449 nan 0.2306 -0.1123
## 3400 22.7664 nan 0.2306 -0.0718
## 3420 22.6948 nan 0.2306 -0.1102
## 3440 22.6788 nan 0.2306 -0.1006
## 3460 22.7039 nan 0.2306 -0.1320
## 3480 22.6540 nan 0.2306 -0.1541
## 3500 22.6096 nan 0.2306 -0.0761
## 3520 22.5975 nan 0.2306 -0.1482
## 3540 22.5583 nan 0.2306 -0.1008
## 3560 22.5509 nan 0.2306 -0.1259
## 3580 22.5035 nan 0.2306 -0.1317
## 3600 22.4361 nan 0.2306 -0.1085
## 3620 22.4121 nan 0.2306 -0.0846
## 3640 22.4368 nan 0.2306 -0.1233
## 3660 22.4090 nan 0.2306 -0.1152
## 3680 22.4103 nan 0.2306 -0.1800
## 3700 22.3624 nan 0.2306 -0.0964
## 3720 22.3273 nan 0.2306 -0.1325
##
## Iter TrainDeviance ValidDeviance StepSize Improve
## 1 677.9177 nan 0.3896 503.0452
## 2 464.4074 nan 0.3896 207.1565
## 3 367.1338 nan 0.3896 94.8912
## 4 316.2130 nan 0.3896 48.2976
## 5 289.9543 nan 0.3896 25.4996
## 6 272.4229 nan 0.3896 17.2176
## 7 261.1223 nan 0.3896 10.0322
## 8 249.5792 nan 0.3896 10.3391
## 9 242.6727 nan 0.3896 6.4213
## 10 237.2319 nan 0.3896 5.2651
## 20 203.0529 nan 0.3896 0.8312
## 40 167.4751 nan 0.3896 0.5045
## 60 151.7658 nan 0.3896 0.4061
## 80 142.7174 nan 0.3896 -0.0960
## 100 131.7149 nan 0.3896 -0.4479
## 120 124.9111 nan 0.3896 -0.1746
## 140 119.4816 nan 0.3896 -0.2282
## 160 113.9785 nan 0.3896 0.0066
## 180 108.4775 nan 0.3896 -0.0310
## 200 105.3022 nan 0.3896 -0.1651
## 220 101.8966 nan 0.3896 0.0017
## 240 98.6545 nan 0.3896 -0.0551
## 260 95.5319 nan 0.3896 0.1276
## 280 92.8357 nan 0.3896 -0.1542
## 300 90.5157 nan 0.3896 0.0413
## 320 88.6628 nan 0.3896 -0.1847
## 340 86.7293 nan 0.3896 -0.0754
## 360 84.5665 nan 0.3896 -0.1408
## 380 82.3831 nan 0.3896 -0.1413
## 400 80.8137 nan 0.3896 -0.1781
## 420 79.3110 nan 0.3896 -0.1592
## 440 77.7796 nan 0.3896 -0.1210
## 460 76.0648 nan 0.3896 -0.1209
## 480 74.8330 nan 0.3896 -0.3387
## 500 73.9394 nan 0.3896 -0.1504
## 520 72.9027 nan 0.3896 -0.0713
## 540 71.6474 nan 0.3896 -0.1716
## 560 70.7949 nan 0.3896 -0.2098
## 580 69.9202 nan 0.3896 -0.1876
## 600 68.6539 nan 0.3896 -0.1238
## 620 67.5875 nan 0.3896 -0.0925
## 640 66.7768 nan 0.3896 -0.1588
## 660 65.9534 nan 0.3896 -0.1482
## 680 65.4190 nan 0.3896 -0.1294
## 700 64.7703 nan 0.3896 -0.0468
## 720 63.8862 nan 0.3896 -0.1381
## 740 63.1045 nan 0.3896 -0.1360
## 760 62.1830 nan 0.3896 -0.1128
## 780 61.5404 nan 0.3896 -0.1327
## 800 60.8238 nan 0.3896 -0.1540
## 820 60.3089 nan 0.3896 -0.0917
## 840 59.7248 nan 0.3896 -0.1272
## 860 59.2553 nan 0.3896 -0.1269
## 880 58.6355 nan 0.3896 -0.1258
## 900 57.8617 nan 0.3896 -0.1275
## 920 57.4151 nan 0.3896 -0.1310
## 940 56.9016 nan 0.3896 -0.0943
## 960 56.2838 nan 0.3896 -0.0737
## 980 55.8246 nan 0.3896 -0.0323
## 1000 55.4061 nan 0.3896 -0.1532
## 1020 54.8312 nan 0.3896 -0.1318
## 1040 54.2491 nan 0.3896 -0.0682
## 1060 53.7525 nan 0.3896 -0.0769
## 1080 53.4079 nan 0.3896 -0.1129
## 1100 53.0612 nan 0.3896 -0.1347
## 1120 52.6730 nan 0.3896 -0.1217
## 1140 52.2650 nan 0.3896 -0.1253
## 1160 51.8112 nan 0.3896 -0.0734
## 1180 51.4796 nan 0.3896 -0.0611
## 1200 51.1854 nan 0.3896 -0.0803
## 1220 50.8764 nan 0.3896 -0.1109
## 1240 50.5035 nan 0.3896 -0.0951
## 1260 50.0948 nan 0.3896 -0.1251
## 1280 49.8616 nan 0.3896 -0.1204
## 1300 49.5378 nan 0.3896 -0.0765
## 1320 49.3096 nan 0.3896 -0.1237
## 1340 49.1052 nan 0.3896 -0.1456
## 1360 48.8655 nan 0.3896 -0.0673
## 1380 48.5866 nan 0.3896 -0.0879
## 1400 48.3641 nan 0.3896 -0.0813
## 1420 48.0880 nan 0.3896 -0.0542
## 1440 47.8252 nan 0.3896 -0.2488
## 1460 47.5270 nan 0.3896 -0.0655
## 1480 47.1983 nan 0.3896 -0.0625
## 1500 46.9476 nan 0.3896 -0.1421
## 1520 46.7118 nan 0.3896 -0.1146
## 1540 46.4980 nan 0.3896 -0.0695
## 1560 46.3585 nan 0.3896 -0.1489
## 1580 46.1005 nan 0.3896 -0.0866
## 1600 45.9274 nan 0.3896 -0.1146
## 1620 45.6893 nan 0.3896 -0.0693
## 1640 45.5171 nan 0.3896 -0.2473
## 1660 45.2703 nan 0.3896 -0.0935
## 1680 44.9753 nan 0.3896 -0.0486
## 1700 44.7373 nan 0.3896 -0.0949
## 1720 44.5174 nan 0.3896 -0.0627
## 1740 44.2596 nan 0.3896 -0.0338
## 1760 44.0570 nan 0.3896 -0.1186
## 1780 43.8737 nan 0.3896 -0.1193
## 1800 43.6546 nan 0.3896 -0.1226
## 1820 43.4366 nan 0.3896 -0.0900
## 1840 43.2918 nan 0.3896 -0.1056
## 1860 43.1475 nan 0.3896 -0.0921
## 1880 42.9396 nan 0.3896 -0.1200
## 1900 42.7686 nan 0.3896 -0.0842
## 1920 42.6692 nan 0.3896 -0.1257
## 1940 42.5785 nan 0.3896 -0.1466
## 1960 42.4136 nan 0.3896 -0.0571
## 1980 42.1970 nan 0.3896 -0.1208
## 2000 42.0065 nan 0.3896 -0.1136
## 2020 41.8058 nan 0.3896 -0.0495
## 2040 41.6862 nan 0.3896 -0.1610
## 2060 41.4721 nan 0.3896 -0.1532
## 2080 41.3383 nan 0.3896 -0.1159
## 2100 41.2395 nan 0.3896 -0.0968
## 2120 41.1132 nan 0.3896 -0.0456
## 2140 40.9744 nan 0.3896 -0.1031
## 2160 40.8291 nan 0.3896 -0.0861
## 2180 40.7293 nan 0.3896 -0.0371
## 2200 40.6236 nan 0.3896 -0.1067
## 2220 40.3859 nan 0.3896 -0.0989
## 2240 40.2353 nan 0.3896 -0.1386
## 2260 40.0893 nan 0.3896 -0.0829
## 2280 39.9074 nan 0.3896 -0.1228
## 2300 39.7626 nan 0.3896 -0.0912
## 2320 39.5553 nan 0.3896 -0.0859
## 2340 39.4032 nan 0.3896 -0.1333
## 2360 39.2632 nan 0.3896 -0.0876
## 2380 39.1619 nan 0.3896 -0.1540
## 2400 39.0062 nan 0.3896 -0.0680
## 2420 38.8738 nan 0.3896 -0.0326
## 2440 38.7165 nan 0.3896 -0.1705
## 2460 38.5941 nan 0.3896 -0.0764
## 2480 38.4791 nan 0.3896 -0.1146
## 2500 38.3235 nan 0.3896 -0.1291
## 2520 38.2607 nan 0.3896 -0.1004
## 2540 38.1519 nan 0.3896 -0.2056
## 2560 38.0500 nan 0.3896 -0.1051
## 2580 37.9584 nan 0.3896 -0.1498
## 2600 37.8551 nan 0.3896 -0.1124
## 2620 37.7084 nan 0.3896 -0.1415
## 2640 37.6301 nan 0.3896 -0.1817
## 2660 37.5307 nan 0.3896 -0.0623
## 2680 37.4721 nan 0.3896 -0.0416
## 2700 37.3431 nan 0.3896 -0.0776
## 2720 37.2182 nan 0.3896 -0.1071
## 2740 37.1750 nan 0.3896 -0.1129
## 2760 37.0388 nan 0.3896 -0.0593
## 2780 36.9265 nan 0.3896 -0.1085
## 2800 36.7457 nan 0.3896 -0.0617
## 2820 36.6568 nan 0.3896 -0.0830
## 2840 36.5762 nan 0.3896 -0.1197
## 2860 36.4910 nan 0.3896 -0.0622
## 2880 36.4077 nan 0.3896 -0.0708
## 2900 36.3286 nan 0.3896 -0.0702
## 2920 36.2666 nan 0.3896 -0.0767
## 2940 36.1699 nan 0.3896 -0.0991
## 2960 36.0668 nan 0.3896 -0.1080
## 2980 35.9766 nan 0.3896 -0.1736
## 3000 35.9440 nan 0.3896 -0.1042
## 3020 35.8180 nan 0.3896 -0.0545
## 3040 35.7321 nan 0.3896 -0.0259
## 3060 35.7266 nan 0.3896 -0.1374
## 3080 35.6597 nan 0.3896 -0.1294
## 3100 35.5163 nan 0.3896 -0.1079
## 3120 35.4309 nan 0.3896 -0.0681
## 3140 35.3129 nan 0.3896 -0.0977
## 3160 35.2468 nan 0.3896 -0.0670
## 3180 35.2879 nan 0.3896 -0.0850
## 3200 35.1955 nan 0.3896 -0.0702
## 3220 35.1500 nan 0.3896 -0.2726
## 3240 34.9397 nan 0.3896 -0.0804
## 3260 34.8954 nan 0.3896 -0.1351
## 3280 34.7431 nan 0.3896 -0.1590
## 3300 34.7291 nan 0.3896 -0.0467
## 3320 34.6046 nan 0.3896 -0.0695
## 3340 34.5626 nan 0.3896 -0.0980
## 3360 34.4333 nan 0.3896 -0.0471
## 3380 34.4596 nan 0.3896 -0.3348
## 3400 34.3071 nan 0.3896 -0.0517
## 3420 34.2082 nan 0.3896 -0.1041
## 3440 34.1743 nan 0.3896 -0.0279
## 3460 34.0891 nan 0.3896 -0.1009
## 3480 33.9961 nan 0.3896 -0.0494
## 3500 33.9499 nan 0.3896 -0.1065
## 3520 33.8591 nan 0.3896 -0.0638
## 3540 33.8026 nan 0.3896 -0.0811
## 3560 33.7864 nan 0.3896 -0.1999
## 3580 33.6565 nan 0.3896 -0.1060
## 3600 33.5864 nan 0.3896 -0.0740
## 3620 33.4546 nan 0.3896 -0.0873
## 3640 33.3822 nan 0.3896 -0.0860
## 3660 33.3406 nan 0.3896 -0.1075
## 3680 33.3233 nan 0.3896 -0.1265
## 3700 33.2399 nan 0.3896 -0.0770
## 3720 33.1669 nan 0.3896 -0.0627
## 3740 33.0661 nan 0.3896 -0.0745
## 3760 33.0417 nan 0.3896 -0.0460
## 3780 33.0633 nan 0.3896 -0.1175
## 3800 32.9575 nan 0.3896 -0.0781
## 3820 32.8900 nan 0.3896 -0.1539
## 3840 32.8047 nan 0.3896 -0.1122
## 3860 32.7430 nan 0.3896 -0.0562
## 3880 32.6737 nan 0.3896 -0.0858
## 3900 32.5778 nan 0.3896 -0.0104
## 3920 32.5213 nan 0.3896 -0.0919
## 3940 32.4966 nan 0.3896 -0.1097
## 3960 32.3621 nan 0.3896 -0.0910
## 3980 32.2821 nan 0.3896 -0.0674
## 4000 32.2862 nan 0.3896 -0.0861
## 4020 32.2191 nan 0.3896 -0.1623
## 4040 32.1737 nan 0.3896 -0.0448
## 4060 32.1100 nan 0.3896 -0.0370
## 4080 32.0268 nan 0.3896 -0.0451
##
## Iter TrainDeviance ValidDeviance StepSize Improve
## 1 545.4580 nan 0.5470 633.6213
## 2 377.1850 nan 0.5470 167.7014
## 3 323.5124 nan 0.5470 52.0036
## 4 292.3818 nan 0.5470 29.8596
## 5 277.8287 nan 0.5470 15.2389
## 6 265.7201 nan 0.5470 10.5742
## 7 258.1538 nan 0.5470 6.7848
## 8 250.2290 nan 0.5470 6.8402
## 9 244.8768 nan 0.5470 4.2257
## 10 237.5742 nan 0.5470 6.5635
## 20 198.3253 nan 0.5470 0.4298
## 40 164.5473 nan 0.5470 0.7642
## 60 148.2691 nan 0.5470 -0.1651
## 80 135.1759 nan 0.5470 -0.1611
## 100 127.8705 nan 0.5470 -0.2108
## 120 121.8714 nan 0.5470 -0.2966
## 140 116.0819 nan 0.5470 -0.2486
## 160 110.9791 nan 0.5470 0.1229
## 180 107.1375 nan 0.5470 -0.1261
## 200 103.5095 nan 0.5470 -0.2410
## 220 99.2128 nan 0.5470 -0.1087
## 240 96.3258 nan 0.5470 0.1416
## 260 93.7276 nan 0.5470 -0.3837
## 280 91.0612 nan 0.5470 -0.1789
## 300 88.5331 nan 0.5470 0.0414
## 320 86.4796 nan 0.5470 -0.2788
## 340 84.6957 nan 0.5470 -0.2004
## 360 83.0994 nan 0.5470 -0.3435
## 380 81.2186 nan 0.5470 -0.0541
## 400 79.3879 nan 0.5470 -0.0995
## 420 78.1239 nan 0.5470 -0.0574
## 440 77.0291 nan 0.5470 -0.1644
## 460 75.6788 nan 0.5470 -0.1747
## 480 74.4667 nan 0.5470 -0.3563
## 500 73.4982 nan 0.5470 -0.0722
## 520 72.4257 nan 0.5470 -0.1029
## 540 71.4823 nan 0.5470 -0.1524
## 560 70.4257 nan 0.5470 -0.3967
## 580 69.2170 nan 0.5470 0.0341
## 600 68.1516 nan 0.5470 -0.2551
## 620 67.1048 nan 0.5470 -0.1916
## 640 66.3775 nan 0.5470 -0.2300
## 660 65.4802 nan 0.5470 -0.0072
## 680 64.6658 nan 0.5470 -0.1738
## 700 63.9307 nan 0.5470 -0.1684
## 720 63.2211 nan 0.5470 -0.2351
## 740 62.5183 nan 0.5470 -0.0749
## 760 61.7917 nan 0.5470 -0.2708
## 780 61.0205 nan 0.5470 -0.3660
## 800 60.4357 nan 0.5470 -0.1108
## 820 59.7303 nan 0.5470 -0.2062
## 840 59.3639 nan 0.5470 -0.2726
## 860 58.8669 nan 0.5470 -0.3277
## 880 58.4731 nan 0.5470 -0.3554
## 900 57.8816 nan 0.5470 -0.0690
## 920 57.2542 nan 0.5470 -0.2311
## 940 56.6573 nan 0.5470 -0.4614
## 960 55.9429 nan 0.5470 -0.1096
## 980 55.5573 nan 0.5470 -0.1985
## 1000 55.1221 nan 0.5470 -0.2684
## 1020 54.5777 nan 0.5470 -0.1457
## 1040 54.1011 nan 0.5470 -0.2034
## 1060 53.6544 nan 0.5470 -0.0960
## 1080 53.3196 nan 0.5470 -0.1556
## 1100 52.9793 nan 0.5470 -0.1098
## 1120 52.6166 nan 0.5470 -0.1734
## 1140 52.0149 nan 0.5470 -0.1783
## 1160 51.8980 nan 0.5470 -0.2139
## 1180 51.2380 nan 0.5470 -0.2300
## 1200 50.6922 nan 0.5470 -0.0717
## 1220 50.3888 nan 0.5470 -0.2266
## 1240 49.8893 nan 0.5470 -0.0659
## 1260 49.6402 nan 0.5470 -0.1468
## 1280 49.3728 nan 0.5470 -0.2729
## 1300 49.1100 nan 0.5470 -0.1799
## 1320 48.6348 nan 0.5470 -0.0082
## 1340 48.2717 nan 0.5470 -0.1372
## 1360 48.1177 nan 0.5470 -0.2228
## 1380 47.8531 nan 0.5470 -0.1912
## 1400 47.6799 nan 0.5470 -0.1968
## 1420 47.4084 nan 0.5470 -0.1358
## 1440 47.0957 nan 0.5470 -0.0161
## 1460 46.9030 nan 0.5470 -0.2450
## 1480 46.6478 nan 0.5470 -0.2480
## 1500 46.2962 nan 0.5470 -0.1306
## 1520 46.0813 nan 0.5470 -0.0039
## 1540 45.9381 nan 0.5470 -0.1601
## 1560 45.7403 nan 0.5470 -0.1269
## 1580 45.7131 nan 0.5470 -0.5127
## 1600 45.4399 nan 0.5470 -0.2599
## 1620 45.3492 nan 0.5470 -0.3068
## 1640 45.0299 nan 0.5470 -0.1249
## 1660 44.7494 nan 0.5470 -0.1197
## 1680 44.5504 nan 0.5470 -0.1773
## 1700 44.4221 nan 0.5470 -0.1089
## 1720 44.2143 nan 0.5470 -0.2453
## 1740 44.0463 nan 0.5470 -0.1090
## 1760 43.9507 nan 0.5470 -0.1771
## 1780 43.7609 nan 0.5470 -0.1044
## 1800 43.5360 nan 0.5470 -0.1517
## 1820 43.3462 nan 0.5470 -0.2261
## 1840 43.1750 nan 0.5470 -0.2599
## 1860 43.1111 nan 0.5470 -0.2407
## 1880 42.8582 nan 0.5470 -0.0957
## 1900 42.7430 nan 0.5470 -0.2706
## 1920 42.6960 nan 0.5470 -0.1439
## 1940 42.5035 nan 0.5470 -0.2001
## 1960 42.2695 nan 0.5470 -0.2848
## 1980 41.9888 nan 0.5470 -0.1946
## 2000 41.8147 nan 0.5470 -0.1361
## 2020 41.7756 nan 0.5470 -0.3177
## 2040 41.6080 nan 0.5470 -0.1187
## 2060 41.5073 nan 0.5470 -0.0446
## 2080 41.3165 nan 0.5470 -0.0402
## 2100 41.2809 nan 0.5470 -0.1384
## 2120 41.1374 nan 0.5470 -0.0157
## 2140 40.8363 nan 0.5470 0.0062
## 2160 40.6662 nan 0.5470 -0.1969
## 2180 40.6572 nan 0.5470 -0.2372
## 2200 40.5961 nan 0.5470 -0.2361
## 2220 40.3117 nan 0.5470 -0.3391
## 2240 40.2932 nan 0.5470 -0.0424
## 2260 40.1921 nan 0.5470 -0.3075
## 2280 40.0839 nan 0.5470 -0.1916
## 2300 39.9071 nan 0.5470 -0.1374
## 2320 39.7727 nan 0.5470 -0.0791
## 2340 39.5200 nan 0.5470 -0.2150
## 2360 39.5038 nan 0.5470 -0.0720
## 2380 39.4039 nan 0.5470 -0.1439
## 2400 39.1900 nan 0.5470 -0.1554
## 2420 39.1232 nan 0.5470 -0.1278
## 2440 39.0439 nan 0.5470 -0.1986
## 2460 38.9547 nan 0.5470 -0.0981
## 2480 38.8996 nan 0.5470 -0.1513
## 2500 38.6764 nan 0.5470 -0.0622
## 2520 38.5672 nan 0.5470 -0.2582
## 2540 38.4975 nan 0.5470 -0.2054
## 2560 38.3661 nan 0.5470 -0.1409
## 2580 38.2186 nan 0.5470 -0.1450
## 2600 38.2406 nan 0.5470 -0.2135
## 2620 38.1130 nan 0.5470 -0.2169
## 2640 38.0224 nan 0.5470 -0.1717
## 2660 37.9413 nan 0.5470 -0.1680
## 2680 37.7852 nan 0.5470 -0.1288
## 2700 37.7605 nan 0.5470 -0.1656
## 2720 37.5747 nan 0.5470 -0.1121
## 2740 37.5366 nan 0.5470 -0.0994
## 2760 37.5062 nan 0.5470 -0.2266
## 2780 37.4467 nan 0.5470 -0.1813
## 2800 37.3650 nan 0.5470 -0.1997
## 2820 37.2200 nan 0.5470 -0.0712
## 2840 37.2164 nan 0.5470 -0.0906
## 2860 37.1314 nan 0.5470 -0.1702
## 2880 37.0179 nan 0.5470 -0.0824
## 2900 36.9730 nan 0.5470 -0.0944
## 2920 36.9626 nan 0.5470 -0.1693
## 2940 36.8526 nan 0.5470 -0.1317
## 2960 36.7494 nan 0.5470 -0.1241
## 2980 36.6386 nan 0.5470 -0.2117
## 3000 36.4818 nan 0.5470 -0.1311
## 3020 36.4702 nan 0.5470 -0.1226
## 3040 36.3716 nan 0.5470 -0.0688
## 3060 36.3829 nan 0.5470 -0.1010
## 3080 36.2960 nan 0.5470 -0.2946
## 3100 36.2825 nan 0.5470 -0.0936
## 3120 36.2139 nan 0.5470 -0.0881
## 3140 35.9991 nan 0.5470 -0.1436
## 3160 35.9321 nan 0.5470 -0.2617
## 3180 35.8465 nan 0.5470 -0.0206
## 3200 35.9516 nan 0.5470 -0.3832
## 3220 35.8693 nan 0.5470 -0.1734
## 3240 35.7836 nan 0.5470 -0.2704
## 3260 35.6468 nan 0.5470 -0.1090
## 3280 35.7024 nan 0.5470 -0.2618
## 3300 35.5157 nan 0.5470 -0.1845
## 3320 35.5796 nan 0.5470 -0.2319
## 3340 35.4250 nan 0.5470 -0.1013
## 3360 35.3250 nan 0.5470 -0.3907
## 3380 35.1500 nan 0.5470 -0.1760
## 3400 35.0896 nan 0.5470 -0.0542
## 3420 35.0693 nan 0.5470 -0.0700
## 3440 34.9707 nan 0.5470 -0.0592
## 3460 34.9264 nan 0.5470 -0.1814
## 3480 34.9231 nan 0.5470 -0.0963
## 3489 34.9143 nan 0.5470 -0.2436
##
## Iter TrainDeviance ValidDeviance StepSize Improve
## 1 817.9795 nan 0.2306 360.4836
## 2 598.4807 nan 0.2306 216.0674
## 3 461.0074 nan 0.2306 137.7353
## 4 374.8007 nan 0.2306 84.3834
## 5 317.1186 nan 0.2306 56.3556
## 6 277.3539 nan 0.2306 39.4376
## 7 247.3365 nan 0.2306 28.4586
## 8 228.7267 nan 0.2306 16.6954
## 9 216.3123 nan 0.2306 11.5026
## 10 206.0406 nan 0.2306 9.6836
## 20 166.0775 nan 0.2306 1.1928
## 40 134.6998 nan 0.2306 0.8234
## 60 118.6869 nan 0.2306 0.2444
## 80 107.4129 nan 0.2306 0.0591
## 100 98.0908 nan 0.2306 0.0024
## 120 91.4347 nan 0.2306 -0.0572
## 140 85.5194 nan 0.2306 -0.1464
## 160 80.8706 nan 0.2306 -0.0789
## 180 76.1076 nan 0.2306 -0.1694
## 200 72.6776 nan 0.2306 -0.1129
## 220 69.0522 nan 0.2306 -0.0469
## 240 66.4354 nan 0.2306 -0.1800
## 260 64.0556 nan 0.2306 -0.1068
## 280 61.4950 nan 0.2306 -0.0726
## 300 59.1289 nan 0.2306 -0.0969
## 320 56.9879 nan 0.2306 -0.2966
## 340 55.3250 nan 0.2306 -0.1203
## 360 53.6108 nan 0.2306 -0.1061
## 380 52.0894 nan 0.2306 -0.0931
## 400 50.8503 nan 0.2306 -0.1582
## 420 49.4024 nan 0.2306 -0.1012
## 440 48.3178 nan 0.2306 -0.2387
## 460 47.1026 nan 0.2306 -0.3775
## 480 46.1321 nan 0.2306 -0.2530
## 500 45.0794 nan 0.2306 -0.1855
## 520 44.1570 nan 0.2306 -0.0929
## 540 43.2673 nan 0.2306 -0.1313
## 560 42.4620 nan 0.2306 -0.1867
## 580 41.8434 nan 0.2306 -0.1059
## 600 41.2033 nan 0.2306 -0.0951
## 620 40.6021 nan 0.2306 -0.1737
## 640 39.9462 nan 0.2306 -0.1408
## 660 39.3580 nan 0.2306 -0.2850
## 680 38.8361 nan 0.2306 -0.0363
## 700 38.3530 nan 0.2306 -0.1149
## 720 37.9192 nan 0.2306 -0.0728
## 740 37.4982 nan 0.2306 -0.2226
## 760 37.0547 nan 0.2306 -0.1346
## 780 36.7159 nan 0.2306 -0.1924
## 800 36.2907 nan 0.2306 -0.1748
## 820 35.9679 nan 0.2306 -0.0885
## 840 35.6256 nan 0.2306 -0.0763
## 860 35.3323 nan 0.2306 -0.2201
## 880 34.9984 nan 0.2306 -0.2433
## 900 34.6927 nan 0.2306 -0.1291
## 920 34.3567 nan 0.2306 -0.0334
## 940 33.9907 nan 0.2306 -0.1187
## 960 33.7249 nan 0.2306 -0.1968
## 980 33.4330 nan 0.2306 -0.1714
## 1000 33.1904 nan 0.2306 -0.1833
## 1020 32.8853 nan 0.2306 -0.1060
## 1040 32.6740 nan 0.2306 -0.1740
## 1060 32.4728 nan 0.2306 -0.0738
## 1080 32.2581 nan 0.2306 -0.1734
## 1100 32.0569 nan 0.2306 -0.1095
## 1120 31.8486 nan 0.2306 -0.0841
## 1140 31.6796 nan 0.2306 -0.2219
## 1160 31.4127 nan 0.2306 -0.1248
## 1180 31.2262 nan 0.2306 -0.1120
## 1200 31.0011 nan 0.2306 -0.0515
## 1220 30.7591 nan 0.2306 -0.0954
## 1240 30.5947 nan 0.2306 -0.1652
## 1260 30.4199 nan 0.2306 -0.0787
## 1280 30.3480 nan 0.2306 -0.1435
## 1300 30.1607 nan 0.2306 -0.1145
## 1320 30.0036 nan 0.2306 -0.1797
## 1340 29.9179 nan 0.2306 -0.1277
## 1360 29.7753 nan 0.2306 -0.1056
## 1380 29.6477 nan 0.2306 -0.3715
## 1400 29.4705 nan 0.2306 -0.1181
## 1420 29.3166 nan 0.2306 -0.1385
## 1440 29.1990 nan 0.2306 -0.0919
## 1460 29.0493 nan 0.2306 -0.0725
## 1480 28.9642 nan 0.2306 -0.2413
## 1500 28.7937 nan 0.2306 -0.1754
## 1520 28.6657 nan 0.2306 -0.1000
## 1540 28.6247 nan 0.2306 -0.0675
## 1560 28.4601 nan 0.2306 -0.1913
## 1580 28.3916 nan 0.2306 -0.0345
## 1600 28.2613 nan 0.2306 -0.1151
## 1620 28.1973 nan 0.2306 -0.1869
## 1640 28.1115 nan 0.2306 -0.1793
## 1660 27.9913 nan 0.2306 -0.0616
## 1680 27.9009 nan 0.2306 -0.0896
## 1700 27.7669 nan 0.2306 -0.1711
## 1720 27.7085 nan 0.2306 -0.1217
## 1740 27.6306 nan 0.2306 -0.1984
## 1760 27.5403 nan 0.2306 -0.1664
## 1780 27.4948 nan 0.2306 -0.0983
## 1800 27.4132 nan 0.2306 -0.1426
## 1820 27.3641 nan 0.2306 -0.1237
## 1840 27.2890 nan 0.2306 -0.1091
## 1860 27.1503 nan 0.2306 -0.1135
## 1880 27.0451 nan 0.2306 -0.1433
## 1900 26.9583 nan 0.2306 -0.1342
## 1920 26.8906 nan 0.2306 -0.0999
## 1940 26.8026 nan 0.2306 -0.0798
## 1960 26.7496 nan 0.2306 -0.1644
## 1980 26.6784 nan 0.2306 -0.1353
## 2000 26.6389 nan 0.2306 -0.1009
## 2020 26.5524 nan 0.2306 -0.0735
## 2040 26.4746 nan 0.2306 -0.1330
## 2060 26.3416 nan 0.2306 -0.0355
## 2080 26.3365 nan 0.2306 -0.1825
## 2100 26.2533 nan 0.2306 -0.1313
## 2120 26.2615 nan 0.2306 -0.2340
## 2140 26.2220 nan 0.2306 -0.1737
## 2160 26.1153 nan 0.2306 -0.1919
## 2180 26.0397 nan 0.2306 -0.1721
## 2200 25.9755 nan 0.2306 -0.1409
## 2220 25.9346 nan 0.2306 -0.0954
## 2240 25.9222 nan 0.2306 -0.1267
## 2260 25.8269 nan 0.2306 -0.1524
## 2280 25.8248 nan 0.2306 -0.1873
## 2300 25.7486 nan 0.2306 -0.0751
## 2320 25.6771 nan 0.2306 -0.0946
## 2340 25.6173 nan 0.2306 -0.1312
## 2360 25.5190 nan 0.2306 -0.0862
## 2380 25.4710 nan 0.2306 -0.1422
## 2400 25.4149 nan 0.2306 -0.2781
## 2420 25.3470 nan 0.2306 -0.0874
## 2440 25.3217 nan 0.2306 -0.1568
## 2460 25.3369 nan 0.2306 -0.1782
## 2480 25.2815 nan 0.2306 -0.1958
## 2500 25.2057 nan 0.2306 -0.1389
## 2520 25.1317 nan 0.2306 -0.1332
## 2540 25.1060 nan 0.2306 -0.1213
## 2560 25.0587 nan 0.2306 -0.0810
## 2580 24.9746 nan 0.2306 -0.0494
## 2600 24.9371 nan 0.2306 -0.1664
## 2620 24.8844 nan 0.2306 -0.1293
## 2640 24.8802 nan 0.2306 -0.1153
## 2660 24.8383 nan 0.2306 -0.0795
## 2680 24.8175 nan 0.2306 -0.1591
## 2700 24.7614 nan 0.2306 -0.2255
## 2720 24.7698 nan 0.2306 -0.1399
## 2740 24.6850 nan 0.2306 -0.1258
## 2760 24.6717 nan 0.2306 -0.0685
## 2780 24.6360 nan 0.2306 -0.1243
## 2800 24.5752 nan 0.2306 -0.1358
## 2820 24.5738 nan 0.2306 -0.0812
## 2840 24.5310 nan 0.2306 -0.1090
## 2860 24.4829 nan 0.2306 -0.1870
## 2880 24.4522 nan 0.2306 -0.0816
## 2900 24.4477 nan 0.2306 -0.1055
## 2920 24.4027 nan 0.2306 -0.0524
## 2940 24.4237 nan 0.2306 -0.2255
## 2960 24.3622 nan 0.2306 -0.1191
## 2980 24.3827 nan 0.2306 -0.2022
## 3000 24.3112 nan 0.2306 -0.0965
## 3020 24.2653 nan 0.2306 -0.1439
## 3040 24.2794 nan 0.2306 -0.1860
## 3060 24.2194 nan 0.2306 -0.2111
## 3080 24.1806 nan 0.2306 -0.0871
## 3100 24.1661 nan 0.2306 -0.1039
## 3120 24.1212 nan 0.2306 -0.1003
## 3140 24.0893 nan 0.2306 -0.1213
## 3160 24.0823 nan 0.2306 -0.0947
## 3180 24.0364 nan 0.2306 -0.1161
## 3200 24.0435 nan 0.2306 -0.0843
## 3220 24.0398 nan 0.2306 -0.1800
## 3240 24.0091 nan 0.2306 -0.1425
## 3260 23.9417 nan 0.2306 -0.2095
## 3280 23.9026 nan 0.2306 -0.1323
## 3300 23.9140 nan 0.2306 -0.1541
## 3320 23.9296 nan 0.2306 -0.1896
## 3340 23.8719 nan 0.2306 -0.1582
## 3360 23.8423 nan 0.2306 -0.1819
## 3380 23.8938 nan 0.2306 -0.1150
## 3400 23.8465 nan 0.2306 -0.1298
## 3420 23.7941 nan 0.2306 -0.0677
## 3440 23.7690 nan 0.2306 -0.1067
## 3460 23.7706 nan 0.2306 -0.1912
## 3480 23.7022 nan 0.2306 -0.1122
## 3500 23.6862 nan 0.2306 -0.1415
## 3520 23.6781 nan 0.2306 -0.1864
## 3540 23.6432 nan 0.2306 -0.1977
## 3560 23.5986 nan 0.2306 -0.1682
## 3580 23.6527 nan 0.2306 -0.1548
## 3600 23.5695 nan 0.2306 -0.1224
## 3620 23.5275 nan 0.2306 -0.0994
## 3640 23.4952 nan 0.2306 -0.1615
## 3660 23.4934 nan 0.2306 -0.1831
## 3680 23.4672 nan 0.2306 -0.1061
## 3700 23.4580 nan 0.2306 -0.1143
## 3720 23.4243 nan 0.2306 -0.1221
##
## Iter TrainDeviance ValidDeviance StepSize Improve
## 1 676.0648 nan 0.3896 501.2356
## 2 468.6518 nan 0.3896 204.2582
## 3 369.5102 nan 0.3896 97.7376
## 4 320.0251 nan 0.3896 46.7955
## 5 293.5891 nan 0.3896 26.6887
## 6 278.3372 nan 0.3896 13.6571
## 7 266.9579 nan 0.3896 10.6189
## 8 256.4446 nan 0.3896 9.7409
## 9 247.6779 nan 0.3896 8.2827
## 10 242.7464 nan 0.3896 3.5754
## 20 205.1517 nan 0.3896 2.4785
## 40 172.9399 nan 0.3896 0.0750
## 60 155.8776 nan 0.3896 0.5588
## 80 144.2854 nan 0.3896 0.0818
## 100 134.4728 nan 0.3896 0.3944
## 120 128.0902 nan 0.3896 -0.0589
## 140 122.6520 nan 0.3896 0.2054
## 160 117.6196 nan 0.3896 0.0489
## 180 113.8429 nan 0.3896 -0.0327
## 200 109.9157 nan 0.3896 -0.1728
## 220 106.2275 nan 0.3896 -0.1117
## 240 103.3688 nan 0.3896 -0.2973
## 260 100.6896 nan 0.3896 -0.1908
## 280 97.8095 nan 0.3896 -0.0845
## 300 95.2219 nan 0.3896 -0.2201
## 320 92.8598 nan 0.3896 -0.3067
## 340 90.6615 nan 0.3896 0.0669
## 360 88.5973 nan 0.3896 -0.0997
## 380 85.9110 nan 0.3896 -0.1100
## 400 84.1636 nan 0.3896 -0.1650
## 420 82.7167 nan 0.3896 -0.1658
## 440 81.4345 nan 0.3896 -0.0900
## 460 79.7939 nan 0.3896 -0.0972
## 480 78.3835 nan 0.3896 -0.2399
## 500 77.0722 nan 0.3896 -0.1633
## 520 75.6129 nan 0.3896 -0.1140
## 540 74.6330 nan 0.3896 -0.0801
## 560 73.5004 nan 0.3896 -0.0417
## 580 72.4234 nan 0.3896 -0.2992
## 600 71.5520 nan 0.3896 -0.1255
## 620 70.7398 nan 0.3896 -0.1498
## 640 69.7438 nan 0.3896 -0.1730
## 660 68.8866 nan 0.3896 -0.0748
## 680 67.9188 nan 0.3896 -0.0662
## 700 66.9143 nan 0.3896 -0.3769
## 720 66.0884 nan 0.3896 -0.1911
## 740 65.3483 nan 0.3896 -0.0906
## 760 64.6753 nan 0.3896 -0.0700
## 780 64.0523 nan 0.3896 -0.1939
## 800 63.5281 nan 0.3896 -0.1176
## 820 62.8711 nan 0.3896 -0.1262
## 840 62.4436 nan 0.3896 -0.1009
## 860 61.8644 nan 0.3896 -0.0923
## 880 61.4104 nan 0.3896 -0.0945
## 900 60.7225 nan 0.3896 -0.1124
## 920 60.0694 nan 0.3896 -0.1315
## 940 59.5196 nan 0.3896 -0.1196
## 960 59.0759 nan 0.3896 -0.0833
## 980 58.5331 nan 0.3896 -0.0986
## 1000 58.1270 nan 0.3896 -0.1328
## 1020 57.6086 nan 0.3896 -0.0817
## 1040 57.1767 nan 0.3896 -0.0751
## 1060 56.6539 nan 0.3896 -0.1598
## 1080 56.3188 nan 0.3896 -0.1946
## 1100 55.6907 nan 0.3896 -0.1118
## 1120 55.3489 nan 0.3896 -0.0897
## 1140 54.8586 nan 0.3896 -0.2510
## 1160 54.5071 nan 0.3896 -0.1355
## 1180 54.2192 nan 0.3896 -0.0890
## 1200 53.8289 nan 0.3896 -0.1476
## 1220 53.4194 nan 0.3896 -0.1035
## 1240 52.8812 nan 0.3896 -0.1733
## 1260 52.3026 nan 0.3896 -0.0509
## 1280 51.8294 nan 0.3896 -0.1012
## 1300 51.4167 nan 0.3896 -0.1613
## 1320 50.9912 nan 0.3896 -0.1227
## 1340 50.5508 nan 0.3896 -0.1409
## 1360 50.2573 nan 0.3896 -0.1109
## 1380 49.9458 nan 0.3896 -0.1292
## 1400 49.6269 nan 0.3896 -0.0875
## 1420 49.4028 nan 0.3896 -0.2602
## 1440 48.9476 nan 0.3896 -0.0634
## 1460 48.7788 nan 0.3896 -0.1279
## 1480 48.5516 nan 0.3896 -0.0972
## 1500 48.1980 nan 0.3896 -0.1070
## 1520 47.9123 nan 0.3896 -0.1049
## 1540 47.5874 nan 0.3896 -0.1224
## 1560 47.3387 nan 0.3896 -0.1435
## 1580 47.1928 nan 0.3896 -0.0440
## 1600 46.8951 nan 0.3896 -0.0656
## 1620 46.7426 nan 0.3896 -0.2407
## 1640 46.4792 nan 0.3896 -0.1516
## 1660 46.1946 nan 0.3896 -0.2040
## 1680 46.0022 nan 0.3896 -0.1520
## 1700 45.7547 nan 0.3896 -0.0729
## 1720 45.4750 nan 0.3896 -0.1421
## 1740 45.2472 nan 0.3896 -0.0800
## 1760 44.9927 nan 0.3896 -0.0260
## 1780 44.8390 nan 0.3896 -0.1538
## 1800 44.6031 nan 0.3896 -0.0304
## 1820 44.3940 nan 0.3896 -0.0709
## 1840 44.2265 nan 0.3896 -0.0992
## 1860 44.0147 nan 0.3896 -0.0784
## 1880 43.8935 nan 0.3896 -0.1505
## 1900 43.6761 nan 0.3896 -0.0577
## 1920 43.5336 nan 0.3896 -0.1672
## 1940 43.3572 nan 0.3896 -0.0822
## 1960 43.0263 nan 0.3896 -0.0485
## 1980 42.8703 nan 0.3896 -0.1239
## 2000 42.7234 nan 0.3896 -0.0567
## 2020 42.6103 nan 0.3896 -0.1466
## 2040 42.5032 nan 0.3896 -0.1370
## 2060 42.3399 nan 0.3896 -0.1311
## 2080 42.1954 nan 0.3896 -0.2589
## 2100 42.0428 nan 0.3896 -0.1091
## 2120 41.8699 nan 0.3896 -0.1300
## 2140 41.7165 nan 0.3896 -0.0816
## 2160 41.6070 nan 0.3896 -0.0827
## 2180 41.4943 nan 0.3896 -0.0760
## 2200 41.3087 nan 0.3896 -0.0846
## 2220 41.0903 nan 0.3896 -0.0855
## 2240 41.0139 nan 0.3896 -0.0868
## 2260 40.8857 nan 0.3896 -0.1006
## 2280 40.7791 nan 0.3896 -0.0956
## 2300 40.6183 nan 0.3896 -0.1300
## 2320 40.4885 nan 0.3896 -0.1643
## 2340 40.4546 nan 0.3896 -0.1659
## 2360 40.2348 nan 0.3896 -0.1000
## 2380 40.0900 nan 0.3896 -0.1117
## 2400 39.8884 nan 0.3896 -0.1246
## 2420 39.8008 nan 0.3896 -0.1558
## 2440 39.6624 nan 0.3896 -0.1886
## 2460 39.6287 nan 0.3896 -0.0925
## 2480 39.4163 nan 0.3896 -0.0781
## 2500 39.4050 nan 0.3896 -0.1879
## 2520 39.3144 nan 0.3896 -0.0885
## 2540 39.1931 nan 0.3896 -0.1413
## 2560 39.0909 nan 0.3896 -0.1362
## 2580 38.9340 nan 0.3896 -0.0596
## 2600 38.8427 nan 0.3896 -0.1551
## 2620 38.7752 nan 0.3896 -0.1435
## 2640 38.5902 nan 0.3896 -0.0676
## 2660 38.4705 nan 0.3896 -0.0938
## 2680 38.3938 nan 0.3896 -0.1119
## 2700 38.3178 nan 0.3896 -0.0753
## 2720 38.2889 nan 0.3896 -0.0718
## 2740 38.1317 nan 0.3896 -0.0923
## 2760 38.0357 nan 0.3896 -0.0048
## 2780 37.9220 nan 0.3896 -0.0881
## 2800 37.8482 nan 0.3896 -0.0761
## 2820 37.7735 nan 0.3896 -0.0737
## 2840 37.7167 nan 0.3896 -0.1507
## 2860 37.6447 nan 0.3896 -0.0714
## 2880 37.5338 nan 0.3896 -0.0626
## 2900 37.4439 nan 0.3896 -0.1703
## 2920 37.3468 nan 0.3896 -0.0818
## 2940 37.1947 nan 0.3896 -0.0695
## 2960 37.1629 nan 0.3896 -0.1228
## 2980 37.0203 nan 0.3896 -0.0807
## 3000 36.9358 nan 0.3896 -0.1167
## 3020 36.8514 nan 0.3896 -0.0725
## 3040 36.7781 nan 0.3896 -0.1615
## 3060 36.7302 nan 0.3896 -0.0852
## 3080 36.6285 nan 0.3896 -0.0410
## 3100 36.4950 nan 0.3896 -0.0870
## 3120 36.3916 nan 0.3896 -0.0630
## 3140 36.3597 nan 0.3896 -0.1348
## 3160 36.2761 nan 0.3896 -0.0961
## 3180 36.1793 nan 0.3896 -0.0695
## 3200 36.0291 nan 0.3896 -0.0845
## 3220 35.9764 nan 0.3896 -0.1373
## 3240 35.8982 nan 0.3896 -0.0676
## 3260 35.7776 nan 0.3896 -0.0993
## 3280 35.7739 nan 0.3896 -0.1754
## 3300 35.7086 nan 0.3896 -0.0697
## 3320 35.6104 nan 0.3896 0.0074
## 3340 35.5593 nan 0.3896 -0.0887
## 3360 35.5020 nan 0.3896 -0.1140
## 3380 35.3907 nan 0.3896 -0.0652
## 3400 35.3092 nan 0.3896 -0.0825
## 3420 35.3100 nan 0.3896 -0.0978
## 3440 35.2294 nan 0.3896 -0.0198
## 3460 35.0980 nan 0.3896 -0.1147
## 3480 35.0338 nan 0.3896 -0.0846
## 3500 35.0240 nan 0.3896 -0.2173
## 3520 35.0554 nan 0.3896 -0.2055
## 3540 34.8980 nan 0.3896 -0.0909
## 3560 34.8176 nan 0.3896 -0.0812
## 3580 34.7459 nan 0.3896 -0.1058
## 3600 34.6967 nan 0.3896 -0.1857
## 3620 34.6003 nan 0.3896 -0.0169
## 3640 34.5725 nan 0.3896 -0.0997
## 3660 34.4810 nan 0.3896 -0.0399
## 3680 34.5286 nan 0.3896 -0.1458
## 3700 34.4055 nan 0.3896 -0.1054
## 3720 34.3393 nan 0.3896 -0.1668
## 3740 34.2928 nan 0.3896 -0.0690
## 3760 34.2423 nan 0.3896 -0.0924
## 3780 34.1979 nan 0.3896 -0.1044
## 3800 34.1073 nan 0.3896 -0.0942
## 3820 34.0990 nan 0.3896 -0.1416
## 3840 34.0286 nan 0.3896 -0.1006
## 3860 33.9801 nan 0.3896 -0.0731
## 3880 33.9531 nan 0.3896 -0.1806
## 3900 33.8676 nan 0.3896 -0.1078
## 3920 33.7365 nan 0.3896 -0.0843
## 3940 33.6904 nan 0.3896 -0.0980
## 3960 33.6399 nan 0.3896 -0.0983
## 3980 33.6576 nan 0.3896 -0.0506
## 4000 33.5392 nan 0.3896 -0.0368
## 4020 33.4918 nan 0.3896 -0.0692
## 4040 33.4557 nan 0.3896 -0.1096
## 4060 33.3201 nan 0.3896 -0.0838
## 4080 33.3455 nan 0.3896 -0.0729
##
## Iter TrainDeviance ValidDeviance StepSize Improve
## 1 541.6083 nan 0.5470 636.1436
## 2 370.0646 nan 0.5470 168.8557
## 3 313.7425 nan 0.5470 54.5681
## 4 285.8516 nan 0.5470 26.6381
## 5 270.1262 nan 0.5470 14.2004
## 6 256.9940 nan 0.5470 13.2382
## 7 247.4635 nan 0.5470 7.2631
## 8 240.0050 nan 0.5470 6.6923
## 9 233.7204 nan 0.5470 5.0025
## 10 225.9685 nan 0.5470 7.2794
## 20 193.2319 nan 0.5470 2.1870
## 40 163.2150 nan 0.5470 0.3537
## 60 146.1770 nan 0.5470 -0.7425
## 80 135.3933 nan 0.5470 -0.5240
## 100 124.2865 nan 0.5470 -0.2836
## 120 118.2953 nan 0.5470 -0.3208
## 140 112.4628 nan 0.5470 0.1064
## 160 107.7171 nan 0.5470 -0.2983
## 180 103.6233 nan 0.5470 -0.2360
## 200 100.1852 nan 0.5470 -0.2237
## 220 96.9017 nan 0.5470 -0.2497
## 240 93.8099 nan 0.5470 -0.1067
## 260 90.9776 nan 0.5470 -0.0615
## 280 88.2163 nan 0.5470 -0.2835
## 300 85.7009 nan 0.5470 -0.2298
## 320 83.9580 nan 0.5470 -0.1897
## 340 82.0331 nan 0.5470 -0.4068
## 360 80.2206 nan 0.5470 -0.2754
## 380 78.7930 nan 0.5470 -0.1792
## 400 77.7421 nan 0.5470 -0.3013
## 420 76.0578 nan 0.5470 -0.2647
## 440 74.8265 nan 0.5470 -0.3330
## 460 73.7321 nan 0.5470 -0.2232
## 480 72.6325 nan 0.5470 -0.1395
## 500 71.4685 nan 0.5470 -0.1924
## 520 70.1102 nan 0.5470 -0.0641
## 540 69.3563 nan 0.5470 -0.1142
## 560 68.2684 nan 0.5470 -0.2694
## 580 67.4660 nan 0.5470 -0.2834
## 600 66.7944 nan 0.5470 -0.3690
## 620 65.7721 nan 0.5470 -0.0915
## 640 64.8431 nan 0.5470 -0.2149
## 660 64.0885 nan 0.5470 -0.0629
## 680 63.2077 nan 0.5470 -0.0675
## 700 62.5557 nan 0.5470 -0.1196
## 720 61.9365 nan 0.5470 -0.0882
## 740 61.5223 nan 0.5470 -0.3251
## 760 60.7210 nan 0.5470 -0.2331
## 780 60.0930 nan 0.5470 -0.1335
## 800 59.5304 nan 0.5470 -0.1128
## 820 59.0534 nan 0.5470 -0.1825
## 840 58.7999 nan 0.5470 -0.1739
## 860 58.4868 nan 0.5470 -0.2244
## 880 57.9725 nan 0.5470 -0.2214
## 900 57.4016 nan 0.5470 -0.3069
## 920 56.7405 nan 0.5470 -0.1607
## 940 56.0748 nan 0.5470 -0.1881
## 960 55.6347 nan 0.5470 -0.2507
## 980 55.1061 nan 0.5470 -0.1446
## 1000 54.6796 nan 0.5470 -0.2802
## 1020 54.2648 nan 0.5470 -0.2905
## 1040 53.8657 nan 0.5470 -0.1711
## 1060 53.3997 nan 0.5470 -0.2236
## 1080 52.8751 nan 0.5470 -0.1636
## 1100 52.6996 nan 0.5470 -0.4565
## 1120 52.3388 nan 0.5470 -0.2636
## 1140 51.9756 nan 0.5470 -0.1102
## 1160 51.5908 nan 0.5470 -0.1930
## 1180 51.1381 nan 0.5470 -0.2381
## 1200 50.9780 nan 0.5470 -0.1290
## 1220 50.4341 nan 0.5470 -0.2706
## 1240 50.1026 nan 0.5470 -0.1748
## 1260 49.7954 nan 0.5470 -0.1334
## 1280 49.5121 nan 0.5470 -0.3638
## 1300 49.2868 nan 0.5470 -0.1547
## 1320 48.9974 nan 0.5470 -0.2079
## 1340 48.6562 nan 0.5470 -0.2062
## 1360 48.3412 nan 0.5470 -0.2585
## 1380 48.0411 nan 0.5470 -0.1828
## 1400 47.7449 nan 0.5470 -0.1628
## 1420 47.5158 nan 0.5470 -0.2113
## 1440 47.2465 nan 0.5470 -0.1194
## 1460 46.8845 nan 0.5470 -0.1525
## 1480 46.6141 nan 0.5470 -0.1405
## 1500 46.5144 nan 0.5470 -0.2423
## 1520 46.3309 nan 0.5470 -0.1459
## 1540 46.2831 nan 0.5470 -1.0062
## 1560 46.1378 nan 0.5470 -0.4530
## 1580 45.8304 nan 0.5470 -0.1792
## 1600 45.6366 nan 0.5470 -0.2370
## 1620 45.3521 nan 0.5470 -0.0818
## 1640 45.1041 nan 0.5470 -0.1231
## 1660 44.8855 nan 0.5470 -0.2310
## 1680 44.5277 nan 0.5470 -0.0878
## 1700 44.2718 nan 0.5470 -0.0579
## 1720 44.1424 nan 0.5470 -0.2465
## 1740 44.0150 nan 0.5470 -0.2961
## 1760 43.8585 nan 0.5470 -0.2561
## 1780 43.7131 nan 0.5470 -0.3731
## 1800 43.4632 nan 0.5470 -0.0801
## 1820 43.4333 nan 0.5470 -0.2216
## 1840 43.2563 nan 0.5470 -0.1434
## 1860 43.1481 nan 0.5470 -0.2650
## 1880 42.9628 nan 0.5470 -0.0593
## 1900 42.6539 nan 0.5470 -0.1161
## 1920 42.5111 nan 0.5470 -0.1866
## 1940 42.3627 nan 0.5470 -0.0603
## 1960 42.2901 nan 0.5470 -0.1565
## 1980 42.1250 nan 0.5470 -0.0666
## 2000 41.9995 nan 0.5470 -0.1283
## 2020 41.8516 nan 0.5470 -0.2224
## 2040 41.6227 nan 0.5470 -0.0353
## 2060 41.5567 nan 0.5470 -0.1596
## 2080 41.2611 nan 0.5470 -0.0057
## 2100 41.2371 nan 0.5470 -0.1911
## 2120 41.2002 nan 0.5470 -0.1603
## 2140 41.1717 nan 0.5470 -0.2339
## 2160 41.0124 nan 0.5470 -0.0785
## 2180 40.8616 nan 0.5470 -0.1504
## 2200 40.7149 nan 0.5470 -0.0667
## 2220 40.5354 nan 0.5470 -0.4046
## 2240 40.4184 nan 0.5470 -0.1981
## 2260 40.2478 nan 0.5470 -0.3158
## 2280 40.0977 nan 0.5470 -0.1932
## 2300 39.9658 nan 0.5470 -0.2602
## 2320 39.8949 nan 0.5470 -0.1310
## 2340 39.6920 nan 0.5470 -0.1954
## 2360 39.5991 nan 0.5470 -0.4759
## 2380 39.4466 nan 0.5470 -0.1083
## 2400 39.3770 nan 0.5470 -0.2319
## 2420 39.2103 nan 0.5470 -0.1624
## 2440 39.0828 nan 0.5470 -0.1379
## 2460 38.9707 nan 0.5470 -0.0916
## 2480 38.9498 nan 0.5470 -0.3318
## 2500 38.6598 nan 0.5470 -0.2023
## 2520 38.5696 nan 0.5470 -0.1426
## 2540 38.4622 nan 0.5470 -0.0932
## 2560 38.2634 nan 0.5470 -0.1747
## 2580 38.2333 nan 0.5470 -0.0531
## 2600 38.2137 nan 0.5470 -0.0999
## 2620 38.1811 nan 0.5470 -0.5628
## 2640 38.0447 nan 0.5470 -0.0684
## 2660 37.9239 nan 0.5470 -0.1771
## 2680 37.9042 nan 0.5470 -0.1053
## 2700 37.8045 nan 0.5470 -0.1121
## 2720 37.7525 nan 0.5470 -0.1981
## 2740 37.6252 nan 0.5470 -0.0256
## 2760 37.5571 nan 0.5470 -0.3739
## 2780 37.4810 nan 0.5470 -0.1618
## 2800 37.3772 nan 0.5470 -0.1383
## 2820 37.3748 nan 0.5470 -0.1305
## 2840 37.3636 nan 0.5470 -0.1884
## 2860 37.1688 nan 0.5470 -0.1112
## 2880 37.0890 nan 0.5470 -0.0930
## 2900 37.0410 nan 0.5470 -0.2062
## 2920 36.8339 nan 0.5470 -0.0909
## 2940 36.7895 nan 0.5470 -0.1916
## 2960 36.6864 nan 0.5470 -0.1507
## 2980 36.6446 nan 0.5470 -0.2444
## 3000 36.7085 nan 0.5470 -0.2183
## 3020 36.6667 nan 0.5470 -0.1746
## 3040 36.5575 nan 0.5470 -0.0687
## 3060 36.5145 nan 0.5470 -0.2578
## 3080 36.4743 nan 0.5470 0.0032
## 3100 36.3920 nan 0.5470 -0.1943
## 3120 36.2956 nan 0.5470 -0.0893
## 3140 36.2251 nan 0.5470 -0.0842
## 3160 36.2301 nan 0.5470 -0.1366
## 3180 36.0991 nan 0.5470 -0.1101
## 3200 35.9998 nan 0.5470 -0.3342
## 3220 35.8326 nan 0.5470 -0.1084
## 3240 35.8030 nan 0.5470 -0.2076
## 3260 35.8218 nan 0.5470 -0.1727
## 3280 35.7564 nan 0.5470 -0.1548
## 3300 35.6639 nan 0.5470 -0.0039
## 3320 35.6029 nan 0.5470 -0.1025
## 3340 35.5631 nan 0.5470 -0.0754
## 3360 35.4512 nan 0.5470 -0.2460
## 3380 35.3859 nan 0.5470 -0.1808
## 3400 35.3564 nan 0.5470 -0.3120
## 3420 35.4314 nan 0.5470 -0.2180
## 3440 35.2765 nan 0.5470 -0.1157
## 3460 35.1668 nan 0.5470 -0.1496
## 3480 35.1188 nan 0.5470 -0.1292
## 3489 35.1927 nan 0.5470 -0.1457
##
## Iter TrainDeviance ValidDeviance StepSize Improve
## 1 824.1791 nan 0.2306 358.2446
## 2 602.5942 nan 0.2306 223.1453
## 3 463.0799 nan 0.2306 135.0700
## 4 374.6483 nan 0.2306 87.0573
## 5 315.7440 nan 0.2306 57.9736
## 6 277.6372 nan 0.2306 37.3516
## 7 251.5559 nan 0.2306 25.0208
## 8 231.8101 nan 0.2306 19.4340
## 9 218.1287 nan 0.2306 12.3669
## 10 206.9774 nan 0.2306 10.6458
## 20 163.5534 nan 0.2306 1.4105
## 40 132.5967 nan 0.2306 0.3350
## 60 117.7354 nan 0.2306 -0.0359
## 80 106.8107 nan 0.2306 -0.1419
## 100 98.5681 nan 0.2306 -0.0179
## 120 90.9076 nan 0.2306 -0.0567
## 140 84.9125 nan 0.2306 -0.2862
## 160 80.1665 nan 0.2306 -0.2086
## 180 75.3929 nan 0.2306 -0.0920
## 200 71.2433 nan 0.2306 0.0180
## 220 67.9380 nan 0.2306 -0.1189
## 240 65.2977 nan 0.2306 0.0098
## 260 62.8858 nan 0.2306 -0.0623
## 280 60.4531 nan 0.2306 -0.0412
## 300 58.3719 nan 0.2306 -0.1843
## 320 56.5066 nan 0.2306 -0.1019
## 340 54.6092 nan 0.2306 -0.1770
## 360 52.9928 nan 0.2306 -0.1207
## 380 51.5212 nan 0.2306 -0.1014
## 400 49.9628 nan 0.2306 -0.1710
## 420 48.7485 nan 0.2306 -0.1628
## 440 47.4420 nan 0.2306 -0.1618
## 460 46.3054 nan 0.2306 -0.0631
## 480 45.3112 nan 0.2306 -0.0089
## 500 44.3890 nan 0.2306 -0.1116
## 520 43.4581 nan 0.2306 -0.2025
## 540 42.6000 nan 0.2306 -0.0255
## 560 41.8766 nan 0.2306 -0.1175
## 580 41.1541 nan 0.2306 -0.0486
## 600 40.4888 nan 0.2306 -0.1108
## 620 39.8739 nan 0.2306 -0.2493
## 640 39.2187 nan 0.2306 -0.0704
## 660 38.5169 nan 0.2306 -0.1071
## 680 37.8931 nan 0.2306 -0.0599
## 700 37.4024 nan 0.2306 -0.1801
## 720 36.8007 nan 0.2306 -0.0723
## 740 36.2326 nan 0.2306 -0.1152
## 760 35.8090 nan 0.2306 -0.2280
## 780 35.3867 nan 0.2306 -0.2246
## 800 34.9869 nan 0.2306 -0.1302
## 820 34.4502 nan 0.2306 -0.1088
## 840 34.1544 nan 0.2306 -0.0968
## 860 33.7443 nan 0.2306 -0.1576
## 880 33.3940 nan 0.2306 -0.1569
## 900 33.1384 nan 0.2306 -0.1269
## 920 32.7725 nan 0.2306 -0.2084
## 940 32.5182 nan 0.2306 -0.1493
## 960 32.2364 nan 0.2306 -0.1101
## 980 31.9601 nan 0.2306 -0.0781
## 1000 31.6693 nan 0.2306 -0.0604
## 1020 31.3130 nan 0.2306 -0.0403
## 1040 31.1118 nan 0.2306 -0.1073
## 1060 30.8807 nan 0.2306 -0.0722
## 1080 30.6953 nan 0.2306 -0.1296
## 1100 30.5683 nan 0.2306 -0.1063
## 1120 30.2876 nan 0.2306 -0.1496
## 1140 30.0851 nan 0.2306 -0.0926
## 1160 29.9268 nan 0.2306 -0.0386
## 1180 29.7517 nan 0.2306 -0.0970
## 1200 29.5675 nan 0.2306 -0.0740
## 1220 29.4328 nan 0.2306 -0.1279
## 1240 29.2892 nan 0.2306 -0.1031
## 1260 29.1074 nan 0.2306 -0.1611
## 1280 28.9510 nan 0.2306 -0.1357
## 1300 28.8129 nan 0.2306 -0.1868
## 1320 28.6876 nan 0.2306 -0.2699
## 1340 28.5318 nan 0.2306 -0.1583
## 1360 28.3877 nan 0.2306 -0.1315
## 1380 28.2421 nan 0.2306 -0.1018
## 1400 28.0664 nan 0.2306 -0.0826
## 1420 27.9392 nan 0.2306 -0.1684
## 1440 27.7951 nan 0.2306 -0.1274
## 1460 27.7004 nan 0.2306 -0.1123
## 1480 27.5709 nan 0.2306 -0.1310
## 1500 27.4284 nan 0.2306 -0.1343
## 1520 27.3364 nan 0.2306 -0.0552
## 1540 27.1638 nan 0.2306 -0.1070
## 1560 27.0503 nan 0.2306 -0.1075
## 1580 26.9914 nan 0.2306 -0.1209
## 1600 26.8289 nan 0.2306 -0.1151
## 1620 26.7256 nan 0.2306 -0.1090
## 1640 26.6368 nan 0.2306 -0.1753
## 1660 26.5661 nan 0.2306 -0.2066
## 1680 26.4189 nan 0.2306 -0.0974
## 1700 26.3229 nan 0.2306 -0.1110
## 1720 26.1898 nan 0.2306 -0.1314
## 1740 26.1169 nan 0.2306 -0.0546
## 1760 26.0244 nan 0.2306 -0.0942
## 1780 25.9758 nan 0.2306 -0.0654
## 1800 25.9276 nan 0.2306 -0.1488
## 1820 25.8199 nan 0.2306 -0.0876
## 1840 25.7263 nan 0.2306 -0.0909
## 1860 25.6422 nan 0.2306 -0.1357
## 1880 25.5902 nan 0.2306 -0.0815
## 1900 25.4785 nan 0.2306 -0.0981
## 1920 25.3365 nan 0.2306 -0.1041
## 1940 25.3208 nan 0.2306 -0.1441
## 1960 25.2230 nan 0.2306 -0.1627
## 1980 25.1191 nan 0.2306 -0.0809
## 2000 25.0568 nan 0.2306 -0.0538
## 2020 25.0206 nan 0.2306 -0.0752
## 2040 25.0037 nan 0.2306 -0.1084
## 2060 24.9322 nan 0.2306 -0.1599
## 2080 24.8867 nan 0.2306 -0.1490
## 2100 24.8245 nan 0.2306 -0.1460
## 2120 24.7587 nan 0.2306 -0.1150
## 2140 24.6998 nan 0.2306 -0.1222
## 2160 24.7084 nan 0.2306 -0.1039
## 2180 24.6559 nan 0.2306 -0.0815
## 2200 24.6070 nan 0.2306 -0.1216
## 2220 24.5250 nan 0.2306 -0.0850
## 2240 24.4679 nan 0.2306 -0.0874
## 2260 24.4272 nan 0.2306 -0.1140
## 2280 24.3430 nan 0.2306 -0.1383
## 2300 24.2886 nan 0.2306 -0.1095
## 2320 24.2214 nan 0.2306 -0.0867
## 2340 24.1815 nan 0.2306 -0.1020
## 2360 24.1820 nan 0.2306 -0.0683
## 2380 24.1362 nan 0.2306 -0.0607
## 2400 24.1275 nan 0.2306 -0.1380
## 2420 24.0435 nan 0.2306 -0.0954
## 2440 24.0182 nan 0.2306 -0.1505
## 2460 23.9880 nan 0.2306 -0.1150
## 2480 23.9417 nan 0.2306 -0.1621
## 2500 23.8582 nan 0.2306 -0.0792
## 2520 23.8636 nan 0.2306 -0.0661
## 2540 23.8234 nan 0.2306 -0.1018
## 2560 23.8074 nan 0.2306 -0.1665
## 2580 23.7704 nan 0.2306 -0.0781
## 2600 23.7107 nan 0.2306 -0.1177
## 2620 23.6841 nan 0.2306 -0.1349
## 2640 23.6538 nan 0.2306 -0.1020
## 2660 23.6234 nan 0.2306 -0.1328
## 2680 23.6007 nan 0.2306 -0.1080
## 2700 23.5655 nan 0.2306 -0.0790
## 2720 23.5933 nan 0.2306 -0.1217
## 2740 23.5512 nan 0.2306 -0.1216
## 2760 23.4836 nan 0.2306 -0.0760
## 2780 23.4662 nan 0.2306 -0.0789
## 2800 23.4250 nan 0.2306 -0.1772
## 2820 23.4035 nan 0.2306 -0.0981
## 2840 23.3607 nan 0.2306 -0.1186
## 2860 23.3284 nan 0.2306 -0.1818
## 2880 23.2648 nan 0.2306 -0.0726
## 2900 23.2493 nan 0.2306 -0.3566
## 2920 23.2352 nan 0.2306 -0.1454
## 2940 23.2144 nan 0.2306 -0.0986
## 2960 23.1937 nan 0.2306 -0.0847
## 2980 23.1873 nan 0.2306 -0.2321
## 3000 23.1565 nan 0.2306 -0.0832
## 3020 23.1238 nan 0.2306 -0.1313
## 3040 23.0693 nan 0.2306 -0.0959
## 3060 23.0845 nan 0.2306 -0.1525
## 3080 23.0624 nan 0.2306 -0.0978
## 3100 23.0172 nan 0.2306 -0.0985
## 3120 22.9917 nan 0.2306 -0.1493
## 3140 22.9472 nan 0.2306 -0.0827
## 3160 22.9172 nan 0.2306 -0.1525
## 3180 22.9283 nan 0.2306 -0.1023
## 3200 22.8934 nan 0.2306 -0.1494
## 3220 22.8558 nan 0.2306 -0.0613
## 3240 22.8180 nan 0.2306 -0.1042
## 3260 22.8153 nan 0.2306 -0.1170
## 3280 22.7372 nan 0.2306 -0.2528
## 3300 22.6994 nan 0.2306 -0.0853
## 3320 22.7109 nan 0.2306 -0.1543
## 3340 22.6615 nan 0.2306 -0.1434
## 3360 22.6107 nan 0.2306 -0.2297
## 3380 22.6079 nan 0.2306 -0.1744
## 3400 22.5634 nan 0.2306 -0.0462
## 3420 22.5848 nan 0.2306 -0.1517
## 3440 22.5851 nan 0.2306 -0.0564
## 3460 22.5221 nan 0.2306 -0.1324
## 3480 22.5139 nan 0.2306 -0.1425
## 3500 22.4823 nan 0.2306 -0.1276
## 3520 22.4176 nan 0.2306 -0.0875
## 3540 22.3976 nan 0.2306 -0.0774
## 3560 22.4229 nan 0.2306 -0.0791
## 3580 22.4124 nan 0.2306 -0.1216
## 3600 22.4153 nan 0.2306 -0.1153
## 3620 22.3811 nan 0.2306 -0.1968
## 3640 22.3235 nan 0.2306 -0.0789
## 3660 22.3481 nan 0.2306 -0.1513
## 3680 22.2957 nan 0.2306 -0.0698
## 3700 22.2775 nan 0.2306 -0.1331
## 3720 22.2341 nan 0.2306 -0.0996
##
## Iter TrainDeviance ValidDeviance StepSize Improve
## 1 671.3874 nan 0.3896 496.5516
## 2 465.5133 nan 0.3896 202.5717
## 3 372.1334 nan 0.3896 91.3579
## 4 321.7179 nan 0.3896 50.1235
## 5 295.4187 nan 0.3896 26.5645
## 6 278.8101 nan 0.3896 15.6311
## 7 266.4982 nan 0.3896 10.6099
## 8 259.1192 nan 0.3896 6.6450
## 9 251.3469 nan 0.3896 6.9884
## 10 245.6628 nan 0.3896 4.8410
## 20 203.1887 nan 0.3896 1.1346
## 40 169.2359 nan 0.3896 0.1007
## 60 152.8039 nan 0.3896 0.2980
## 80 143.0199 nan 0.3896 -0.2587
## 100 133.9578 nan 0.3896 -0.0136
## 120 127.3904 nan 0.3896 -0.3946
## 140 120.8181 nan 0.3896 0.1822
## 160 115.7616 nan 0.3896 0.0565
## 180 111.9011 nan 0.3896 -0.1597
## 200 108.2999 nan 0.3896 -0.0946
## 220 104.3543 nan 0.3896 0.0802
## 240 100.9133 nan 0.3896 -0.0854
## 260 98.1068 nan 0.3896 -0.2004
## 280 95.3875 nan 0.3896 0.0962
## 300 92.7738 nan 0.3896 -0.2250
## 320 90.7452 nan 0.3896 -0.1845
## 340 88.5841 nan 0.3896 0.0973
## 360 86.8098 nan 0.3896 -0.1892
## 380 85.0092 nan 0.3896 -0.1039
## 400 82.8555 nan 0.3896 0.0283
## 420 80.9717 nan 0.3896 -0.1535
## 440 79.2532 nan 0.3896 -0.0257
## 460 78.0586 nan 0.3896 -0.1740
## 480 76.7350 nan 0.3896 0.0409
## 500 75.1397 nan 0.3896 -0.1287
## 520 73.7946 nan 0.3896 -0.1500
## 540 72.4618 nan 0.3896 -0.0033
## 560 71.4492 nan 0.3896 -0.1380
## 580 70.4239 nan 0.3896 -0.0673
## 600 69.3448 nan 0.3896 -0.1040
## 620 68.3892 nan 0.3896 -0.1477
## 640 67.5616 nan 0.3896 -0.1621
## 660 66.4883 nan 0.3896 -0.1535
## 680 65.6748 nan 0.3896 -0.1677
## 700 64.6699 nan 0.3896 -0.0671
## 720 64.0246 nan 0.3896 -0.1522
## 740 63.2391 nan 0.3896 -0.1262
## 760 62.5489 nan 0.3896 -0.0833
## 780 61.8850 nan 0.3896 -0.1243
## 800 61.0850 nan 0.3896 -0.0690
## 820 60.3331 nan 0.3896 -0.0793
## 840 59.7439 nan 0.3896 -0.1801
## 860 59.0730 nan 0.3896 -0.0725
## 880 58.4614 nan 0.3896 -0.1300
## 900 57.8612 nan 0.3896 -0.0860
## 920 57.3315 nan 0.3896 -0.0956
## 940 56.6954 nan 0.3896 -0.0425
## 960 56.3049 nan 0.3896 -0.0876
## 980 55.7520 nan 0.3896 -0.1025
## 1000 55.0975 nan 0.3896 -0.0285
## 1020 54.4050 nan 0.3896 -0.1001
## 1040 53.8823 nan 0.3896 -0.0737
## 1060 53.1902 nan 0.3896 -0.0110
## 1080 52.7352 nan 0.3896 -0.1193
## 1100 52.2175 nan 0.3896 -0.0304
## 1120 51.8813 nan 0.3896 -0.2479
## 1140 51.6212 nan 0.3896 -0.1871
## 1160 51.1417 nan 0.3896 -0.0622
## 1180 50.5718 nan 0.3896 -0.0758
## 1200 50.2071 nan 0.3896 -0.0858
## 1220 49.7801 nan 0.3896 -0.0833
## 1240 49.2871 nan 0.3896 -0.0514
## 1260 48.9010 nan 0.3896 -0.0268
## 1280 48.5616 nan 0.3896 -0.0993
## 1300 48.1681 nan 0.3896 -0.0627
## 1320 47.9208 nan 0.3896 -0.0930
## 1340 47.6392 nan 0.3896 -0.0530
## 1360 47.3173 nan 0.3896 -0.1755
## 1380 46.9239 nan 0.3896 -0.1102
## 1400 46.7107 nan 0.3896 -0.2095
## 1420 46.3897 nan 0.3896 -0.0147
## 1440 46.1238 nan 0.3896 -0.0796
## 1460 45.8396 nan 0.3896 -0.1549
## 1480 45.5258 nan 0.3896 -0.0945
## 1500 45.2813 nan 0.3896 -0.1546
## 1520 44.9238 nan 0.3896 -0.0774
## 1540 44.6608 nan 0.3896 -0.0927
## 1560 44.4247 nan 0.3896 -0.1341
## 1580 44.0255 nan 0.3896 -0.1895
## 1600 43.8396 nan 0.3896 -0.0411
## 1620 43.6158 nan 0.3896 -0.1008
## 1640 43.2151 nan 0.3896 -0.0889
## 1660 42.9304 nan 0.3896 -0.0672
## 1680 42.7607 nan 0.3896 -0.0851
## 1700 42.4814 nan 0.3896 -0.0638
## 1720 42.2849 nan 0.3896 -0.1624
## 1740 42.0428 nan 0.3896 -0.0703
## 1760 41.8265 nan 0.3896 -0.0763
## 1780 41.6855 nan 0.3896 -0.0805
## 1800 41.5593 nan 0.3896 -0.0951
## 1820 41.3660 nan 0.3896 -0.0593
## 1840 41.1854 nan 0.3896 -0.0461
## 1860 40.9923 nan 0.3896 -0.0796
## 1880 40.7317 nan 0.3896 -0.1099
## 1900 40.6201 nan 0.3896 -0.1279
## 1920 40.4110 nan 0.3896 -0.1118
## 1940 40.1738 nan 0.3896 -0.0382
## 1960 40.0084 nan 0.3896 -0.0840
## 1980 39.8863 nan 0.3896 -0.0691
## 2000 39.7297 nan 0.3896 -0.0531
## 2020 39.5852 nan 0.3896 -0.0746
## 2040 39.3711 nan 0.3896 -0.0334
## 2060 39.1950 nan 0.3896 -0.0687
## 2080 39.0210 nan 0.3896 -0.0093
## 2100 38.8704 nan 0.3896 -0.0531
## 2120 38.7267 nan 0.3896 -0.2325
## 2140 38.4904 nan 0.3896 -0.0961
## 2160 38.3543 nan 0.3896 -0.1334
## 2180 38.1819 nan 0.3896 -0.0501
## 2200 38.0542 nan 0.3896 -0.1355
## 2220 37.9045 nan 0.3896 -0.1364
## 2240 37.7468 nan 0.3896 -0.1044
## 2260 37.6469 nan 0.3896 -0.0837
## 2280 37.5027 nan 0.3896 -0.0477
## 2300 37.4376 nan 0.3896 -0.0450
## 2320 37.3645 nan 0.3896 -0.0802
## 2340 37.2505 nan 0.3896 -0.0754
## 2360 37.0862 nan 0.3896 -0.1140
## 2380 36.9572 nan 0.3896 -0.0871
## 2400 36.8396 nan 0.3896 -0.1054
## 2420 36.7218 nan 0.3896 -0.1447
## 2440 36.5908 nan 0.3896 -0.0552
## 2460 36.4604 nan 0.3896 -0.0442
## 2480 36.2730 nan 0.3896 -0.0181
## 2500 36.1565 nan 0.3896 -0.0565
## 2520 36.0297 nan 0.3896 -0.0726
## 2540 35.8727 nan 0.3896 -0.0707
## 2560 35.7187 nan 0.3896 -0.0682
## 2580 35.6300 nan 0.3896 -0.0754
## 2600 35.5242 nan 0.3896 -0.0609
## 2620 35.4421 nan 0.3896 -0.0327
## 2640 35.3000 nan 0.3896 -0.0717
## 2660 35.2187 nan 0.3896 -0.0719
## 2680 35.1711 nan 0.3896 -0.0943
## 2700 35.1312 nan 0.3896 -0.0623
## 2720 35.0422 nan 0.3896 -0.0661
## 2740 34.9719 nan 0.3896 -0.0945
## 2760 34.8904 nan 0.3896 -0.0697
## 2780 34.7694 nan 0.3896 -0.0628
## 2800 34.6740 nan 0.3896 -0.0287
## 2820 34.6352 nan 0.3896 -0.0574
## 2840 34.5310 nan 0.3896 -0.0655
## 2860 34.4333 nan 0.3896 -0.0444
## 2880 34.3222 nan 0.3896 -0.0866
## 2900 34.3011 nan 0.3896 -0.0609
## 2920 34.1628 nan 0.3896 -0.0650
## 2940 34.0965 nan 0.3896 -0.0597
## 2960 34.0566 nan 0.3896 -0.0172
## 2980 33.9456 nan 0.3896 -0.0455
## 3000 33.8259 nan 0.3896 -0.0609
## 3020 33.7196 nan 0.3896 -0.0741
## 3040 33.6119 nan 0.3896 -0.0579
## 3060 33.5874 nan 0.3896 -0.1106
## 3080 33.4885 nan 0.3896 -0.0794
## 3100 33.4744 nan 0.3896 -0.1849
## 3120 33.3023 nan 0.3896 -0.0478
## 3140 33.2205 nan 0.3896 -0.0928
## 3160 33.1342 nan 0.3896 -0.0890
## 3180 33.0328 nan 0.3896 -0.0536
## 3200 32.9577 nan 0.3896 -0.0533
## 3220 32.8530 nan 0.3896 -0.0681
## 3240 32.8654 nan 0.3896 -0.0471
## 3260 32.7768 nan 0.3896 -0.0495
## 3280 32.7077 nan 0.3896 -0.0832
## 3300 32.6538 nan 0.3896 -0.0594
## 3320 32.5893 nan 0.3896 -0.0661
## 3340 32.5311 nan 0.3896 -0.0406
## 3360 32.5570 nan 0.3896 -0.0441
## 3380 32.4354 nan 0.3896 -0.0945
## 3400 32.3727 nan 0.3896 -0.0512
## 3420 32.3238 nan 0.3896 -0.0316
## 3440 32.2384 nan 0.3896 -0.0697
## 3460 32.1871 nan 0.3896 -0.0429
## 3480 32.0972 nan 0.3896 -0.0794
## 3500 31.9854 nan 0.3896 -0.0442
## 3520 31.9280 nan 0.3896 -0.0448
## 3540 31.8480 nan 0.3896 -0.0741
## 3560 31.8263 nan 0.3896 -0.1445
## 3580 31.7264 nan 0.3896 -0.0686
## 3600 31.6222 nan 0.3896 -0.0887
## 3620 31.6219 nan 0.3896 -0.1684
## 3640 31.5192 nan 0.3896 -0.0788
## 3660 31.4653 nan 0.3896 -0.0824
## 3680 31.3744 nan 0.3896 -0.1077
## 3700 31.3040 nan 0.3896 -0.0443
## 3720 31.3153 nan 0.3896 -0.1447
## 3740 31.2481 nan 0.3896 -0.0601
## 3760 31.1790 nan 0.3896 -0.0607
## 3780 31.2181 nan 0.3896 -0.5130
## 3800 31.1134 nan 0.3896 -0.0753
## 3820 30.9845 nan 0.3896 -0.0740
## 3840 30.9393 nan 0.3896 -0.0708
## 3860 30.9222 nan 0.3896 -0.0265
## 3880 30.8419 nan 0.3896 -0.0291
## 3900 30.8354 nan 0.3896 -0.0721
## 3920 30.7764 nan 0.3896 -0.1259
## 3940 30.7419 nan 0.3896 -0.1150
## 3960 30.6856 nan 0.3896 -0.0721
## 3980 30.6447 nan 0.3896 -0.0417
## 4000 30.5744 nan 0.3896 -0.0671
## 4020 30.5236 nan 0.3896 -0.0719
## 4040 30.4486 nan 0.3896 -0.0713
## 4060 30.4135 nan 0.3896 -0.0753
## 4080 30.3779 nan 0.3896 -0.1855
##
## Iter TrainDeviance ValidDeviance StepSize Improve
## 1 538.9561 nan 0.5470 636.7305
## 2 372.8056 nan 0.5470 169.2664
## 3 313.4867 nan 0.5470 57.0167
## 4 284.6830 nan 0.5470 26.5043
## 5 270.2090 nan 0.5470 14.4109
## 6 258.7196 nan 0.5470 11.0110
## 7 249.9269 nan 0.5470 7.2169
## 8 239.4164 nan 0.5470 8.9650
## 9 231.3752 nan 0.5470 6.9971
## 10 225.2811 nan 0.5470 5.1051
## 20 188.4264 nan 0.5470 2.1321
## 40 160.3528 nan 0.5470 0.7179
## 60 144.8292 nan 0.5470 0.0354
## 80 136.0276 nan 0.5470 -0.3044
## 100 125.8924 nan 0.5470 -0.1985
## 120 117.9102 nan 0.5470 0.0537
## 140 112.2190 nan 0.5470 -0.4669
## 160 106.8878 nan 0.5470 -0.1133
## 180 103.2165 nan 0.5470 -0.1942
## 200 100.2301 nan 0.5470 -0.4626
## 220 96.6519 nan 0.5470 0.0180
## 240 93.7008 nan 0.5470 -0.2279
## 260 90.4649 nan 0.5470 -0.0165
## 280 87.6059 nan 0.5470 -0.1844
## 300 85.5410 nan 0.5470 -0.1157
## 320 83.1453 nan 0.5470 -0.1128
## 340 81.1772 nan 0.5470 -0.1831
## 360 79.4130 nan 0.5470 -0.2509
## 380 77.6551 nan 0.5470 0.0630
## 400 76.1642 nan 0.5470 -0.2696
## 420 74.3502 nan 0.5470 -0.2899
## 440 73.2944 nan 0.5470 -0.3182
## 460 71.8284 nan 0.5470 -0.0697
## 480 70.1515 nan 0.5470 -0.1457
## 500 69.1432 nan 0.5470 -0.1301
## 520 67.7831 nan 0.5470 -0.1493
## 540 66.7442 nan 0.5470 -0.2217
## 560 65.9687 nan 0.5470 -0.1190
## 580 65.2866 nan 0.5470 -0.8582
## 600 63.9886 nan 0.5470 -0.0942
## 620 63.1216 nan 0.5470 -0.2450
## 640 62.1326 nan 0.5470 -0.2753
## 660 61.2606 nan 0.5470 -0.1086
## 680 60.4858 nan 0.5470 -0.2737
## 700 59.6035 nan 0.5470 -0.1940
## 720 58.8562 nan 0.5470 -0.1260
## 740 58.0577 nan 0.5470 -0.3069
## 760 57.5289 nan 0.5470 -0.3477
## 780 56.6959 nan 0.5470 -0.0680
## 800 56.1329 nan 0.5470 -0.1153
## 820 55.4822 nan 0.5470 -0.1391
## 840 54.8053 nan 0.5470 -0.1702
## 860 53.9178 nan 0.5470 -0.0820
## 880 53.5737 nan 0.5470 -0.0318
## 900 53.0706 nan 0.5470 -0.2020
## 920 52.6331 nan 0.5470 -0.1151
## 940 52.2214 nan 0.5470 -0.1514
## 960 51.6577 nan 0.5470 -0.2383
## 980 50.9177 nan 0.5470 -0.2011
## 1000 50.3114 nan 0.5470 -0.1761
## 1020 50.0623 nan 0.5470 -0.0690
## 1040 49.4910 nan 0.5470 -0.0865
## 1060 49.0702 nan 0.5470 -0.1771
## 1080 48.6451 nan 0.5470 -0.1555
## 1100 48.2811 nan 0.5470 -0.0687
## 1120 47.9946 nan 0.5470 -0.2420
## 1140 47.5489 nan 0.5470 -0.0638
## 1160 47.1475 nan 0.5470 -0.1326
## 1180 46.8413 nan 0.5470 -0.1945
## 1200 46.3953 nan 0.5470 -0.1621
## 1220 46.0213 nan 0.5470 -0.1926
## 1240 45.6132 nan 0.5470 -0.2429
## 1260 45.3134 nan 0.5470 -0.0775
## 1280 45.0545 nan 0.5470 -0.0866
## 1300 44.7921 nan 0.5470 -0.1168
## 1320 44.3905 nan 0.5470 -0.1146
## 1340 43.9337 nan 0.5470 -0.1765
## 1360 43.7999 nan 0.5470 -0.1624
## 1380 43.3672 nan 0.5470 -0.1511
## 1400 43.1658 nan 0.5470 -0.0998
## 1420 43.0283 nan 0.5470 -0.1919
## 1440 42.8793 nan 0.5470 -0.1314
## 1460 42.6555 nan 0.5470 -0.1814
## 1480 42.4577 nan 0.5470 -0.0925
## 1500 42.1597 nan 0.5470 -0.0928
## 1520 42.0237 nan 0.5470 -0.1920
## 1540 41.7537 nan 0.5470 -0.0323
## 1560 41.5616 nan 0.5470 -0.1722
## 1580 41.3772 nan 0.5470 -0.0436
## 1600 41.1877 nan 0.5470 -0.0736
## 1620 40.9182 nan 0.5470 -0.0917
## 1640 40.8450 nan 0.5470 -0.0878
## 1660 40.4850 nan 0.5470 -0.1034
## 1680 40.3796 nan 0.5470 -0.1729
## 1700 40.0781 nan 0.5470 -0.0295
## 1720 39.8253 nan 0.5470 -0.0903
## 1740 39.6966 nan 0.5470 -0.1137
## 1760 39.4600 nan 0.5470 -0.1501
## 1780 39.4242 nan 0.5470 -0.0625
## 1800 39.2284 nan 0.5470 -0.0951
## 1820 39.0761 nan 0.5470 -0.2690
## 1840 38.9349 nan 0.5470 -0.0929
## 1860 38.8005 nan 0.5470 -0.1060
## 1880 38.6005 nan 0.5470 -0.0999
## 1900 38.4188 nan 0.5470 -0.1228
## 1920 38.3166 nan 0.5470 -0.2488
## 1940 38.2092 nan 0.5470 -0.3127
## 1960 37.9593 nan 0.5470 -0.0577
## 1980 37.8165 nan 0.5470 -0.0970
## 2000 37.7148 nan 0.5470 -0.2030
## 2020 37.5467 nan 0.5470 -0.1078
## 2040 37.5087 nan 0.5470 -0.1176
## 2060 37.3156 nan 0.5470 -0.1616
## 2080 37.2686 nan 0.5470 -0.1534
## 2100 37.0329 nan 0.5470 -0.2425
## 2120 36.8316 nan 0.5470 -0.0462
## 2140 36.8374 nan 0.5470 -0.0858
## 2160 36.7047 nan 0.5470 -0.1194
## 2180 36.6385 nan 0.5470 -0.2225
## 2200 36.5006 nan 0.5470 -0.1453
## 2220 36.3954 nan 0.5470 -0.1459
## 2240 36.4256 nan 0.5470 -0.2324
## 2260 36.3313 nan 0.5470 -0.1100
## 2280 36.0971 nan 0.5470 -0.1543
## 2300 36.0353 nan 0.5470 -0.0260
## 2320 35.9727 nan 0.5470 -0.0617
## 2340 35.9430 nan 0.5470 -0.1435
## 2360 35.8695 nan 0.5470 -0.1958
## 2380 35.6545 nan 0.5470 -0.1495
## 2400 35.5225 nan 0.5470 -0.0909
## 2420 35.4419 nan 0.5470 -0.0932
## 2440 35.3119 nan 0.5470 -0.1483
## 2460 35.1598 nan 0.5470 -0.0995
## 2480 35.0076 nan 0.5470 -0.0703
## 2500 34.9642 nan 0.5470 -0.1402
## 2520 34.9331 nan 0.5470 -0.0457
## 2540 34.7587 nan 0.5470 -0.0977
## 2560 34.6945 nan 0.5470 -0.0517
## 2580 34.5330 nan 0.5470 -0.1000
## 2600 34.5991 nan 0.5470 -0.1824
## 2620 34.4184 nan 0.5470 -0.1209
## 2640 34.3530 nan 0.5470 -0.1355
## 2660 34.1998 nan 0.5470 -0.0546
## 2680 34.0698 nan 0.5470 -0.0940
## 2700 33.9838 nan 0.5470 -0.0425
## 2720 33.8366 nan 0.5470 -0.1449
## 2740 33.8005 nan 0.5470 -0.1071
## 2760 33.7960 nan 0.5470 -0.1808
## 2780 33.7187 nan 0.5470 -0.0380
## 2800 33.6996 nan 0.5470 -0.2309
## 2820 33.5618 nan 0.5470 -0.1126
## 2840 33.6004 nan 0.5470 -0.0810
## 2860 33.5039 nan 0.5470 -0.0750
## 2880 33.4186 nan 0.5470 -0.1165
## 2900 33.4167 nan 0.5470 -0.1119
## 2920 33.3660 nan 0.5470 -0.0688
## 2940 33.1700 nan 0.5470 -0.1216
## 2960 33.0975 nan 0.5470 -0.0861
## 2980 33.0973 nan 0.5470 -0.2253
## 3000 33.1051 nan 0.5470 -0.1028
## 3020 33.0100 nan 0.5470 -0.0472
## 3040 32.9987 nan 0.5470 -0.1237
## 3060 32.8058 nan 0.5470 -0.0533
## 3080 32.7599 nan 0.5470 -0.0551
## 3100 32.6907 nan 0.5470 -0.0401
## 3120 32.6287 nan 0.5470 -0.1362
## 3140 32.5585 nan 0.5470 -0.1315
## 3160 32.5624 nan 0.5470 -0.1241
## 3180 32.4214 nan 0.5470 -0.1756
## 3200 32.3856 nan 0.5470 -0.1198
## 3220 32.4105 nan 0.5470 -0.0655
## 3240 32.4140 nan 0.5470 -0.1539
## 3260 32.3337 nan 0.5470 -0.1326
## 3280 32.3275 nan 0.5470 -0.0801
## 3300 32.2570 nan 0.5470 -0.0442
## 3320 32.2882 nan 0.5470 -0.0528
## 3340 32.2261 nan 0.5470 -0.1130
## 3360 32.0559 nan 0.5470 -0.0406
## 3380 32.0355 nan 0.5470 -0.2084
## 3400 31.9923 nan 0.5470 -0.1296
## 3420 31.9452 nan 0.5470 -0.2047
## 3440 31.8718 nan 0.5470 -0.1210
## 3460 31.8225 nan 0.5470 -0.0629
## 3480 31.8046 nan 0.5470 -0.1712
## 3489 31.7651 nan 0.5470 -0.0707
##
## Iter TrainDeviance ValidDeviance StepSize Improve
## 1 821.2735 nan 0.2306 357.4904
## 2 604.8452 nan 0.2306 215.4831
## 3 463.5180 nan 0.2306 139.3365
## 4 373.6059 nan 0.2306 87.7389
## 5 316.4971 nan 0.2306 54.6248
## 6 277.7644 nan 0.2306 38.9442
## 7 250.4409 nan 0.2306 25.8902
## 8 232.1320 nan 0.2306 17.4474
## 9 216.2898 nan 0.2306 14.0407
## 10 205.7312 nan 0.2306 10.1643
## 20 162.9928 nan 0.2306 1.5304
## 40 132.5189 nan 0.2306 0.9219
## 60 116.8966 nan 0.2306 0.2388
## 80 105.4048 nan 0.2306 -0.0405
## 100 96.6068 nan 0.2306 0.0226
## 120 89.6807 nan 0.2306 0.0752
## 140 83.0713 nan 0.2306 -0.2003
## 160 77.9238 nan 0.2306 -0.1739
## 180 74.4981 nan 0.2306 -0.1720
## 200 71.1431 nan 0.2306 -0.0590
## 220 67.7049 nan 0.2306 -0.1382
## 240 65.0286 nan 0.2306 -0.1712
## 260 62.1297 nan 0.2306 -0.1159
## 280 59.8787 nan 0.2306 0.0329
## 300 58.1236 nan 0.2306 -0.0487
## 320 56.1503 nan 0.2306 -0.0887
## 340 54.5271 nan 0.2306 -0.0868
## 360 52.9863 nan 0.2306 -0.1056
## 380 51.5269 nan 0.2306 -0.1374
## 400 50.2964 nan 0.2306 -0.1628
## 420 49.1816 nan 0.2306 -0.0944
## 440 47.9502 nan 0.2306 -0.1871
## 460 46.9319 nan 0.2306 -0.2246
## 480 45.8832 nan 0.2306 -0.1970
## 500 44.9356 nan 0.2306 -0.2268
## 520 44.0038 nan 0.2306 -0.1062
## 540 43.1642 nan 0.2306 -0.0839
## 560 42.3752 nan 0.2306 -0.1416
## 580 41.7386 nan 0.2306 -0.1176
## 600 41.0385 nan 0.2306 -0.1092
## 620 40.3456 nan 0.2306 -0.1081
## 640 39.8137 nan 0.2306 -0.2815
## 660 39.1114 nan 0.2306 -0.0619
## 680 38.5531 nan 0.2306 -0.0913
## 700 37.9400 nan 0.2306 -0.0480
## 720 37.4194 nan 0.2306 -0.1096
## 740 36.8694 nan 0.2306 -0.1480
## 760 36.4566 nan 0.2306 -0.0110
## 780 35.9858 nan 0.2306 -0.1092
## 800 35.5873 nan 0.2306 -0.1181
## 820 35.2374 nan 0.2306 -0.0806
## 840 34.9362 nan 0.2306 -0.1044
## 860 34.5834 nan 0.2306 -0.1046
## 880 34.2821 nan 0.2306 -0.0581
## 900 33.9424 nan 0.2306 -0.1737
## 920 33.6521 nan 0.2306 -0.1606
## 940 33.4270 nan 0.2306 -0.2466
## 960 33.1245 nan 0.2306 -0.1591
## 980 32.8198 nan 0.2306 -0.1089
## 1000 32.5575 nan 0.2306 -0.0237
## 1020 32.2492 nan 0.2306 -0.1652
## 1040 32.0562 nan 0.2306 -0.1987
## 1060 31.8081 nan 0.2306 -0.0816
## 1080 31.6079 nan 0.2306 -0.1078
## 1100 31.4072 nan 0.2306 -0.1677
## 1120 31.2550 nan 0.2306 -0.1665
## 1140 31.0187 nan 0.2306 -0.0917
## 1160 30.7679 nan 0.2306 -0.1561
## 1180 30.5728 nan 0.2306 -0.1489
## 1200 30.3817 nan 0.2306 -0.1030
## 1220 30.2398 nan 0.2306 -0.2279
## 1240 30.0670 nan 0.2306 -0.1959
## 1260 29.8925 nan 0.2306 -0.1366
## 1280 29.7273 nan 0.2306 -0.0874
## 1300 29.5735 nan 0.2306 -0.1220
## 1320 29.4295 nan 0.2306 -0.1060
## 1340 29.2611 nan 0.2306 -0.1572
## 1360 29.1115 nan 0.2306 -0.1770
## 1380 28.9512 nan 0.2306 -0.1768
## 1400 28.8648 nan 0.2306 -0.0915
## 1420 28.7821 nan 0.2306 -0.1221
## 1440 28.6402 nan 0.2306 -0.1663
## 1460 28.4590 nan 0.2306 -0.1337
## 1480 28.3540 nan 0.2306 -0.0838
## 1500 28.2011 nan 0.2306 -0.1187
## 1520 28.0930 nan 0.2306 -0.0584
## 1540 28.0092 nan 0.2306 -0.2302
## 1560 27.8745 nan 0.2306 -0.1271
## 1580 27.8177 nan 0.2306 -0.0767
## 1600 27.6988 nan 0.2306 -0.1260
## 1620 27.5336 nan 0.2306 -0.0976
## 1640 27.4670 nan 0.2306 -0.1018
## 1660 27.3787 nan 0.2306 -0.1415
## 1680 27.3251 nan 0.2306 -0.1669
## 1700 27.2459 nan 0.2306 -0.1546
## 1720 27.1807 nan 0.2306 -0.0621
## 1740 27.1511 nan 0.2306 -0.1601
## 1760 27.0252 nan 0.2306 -0.0846
## 1780 26.9515 nan 0.2306 -0.1547
## 1800 26.8762 nan 0.2306 -0.1122
## 1820 26.8005 nan 0.2306 -0.1037
## 1840 26.7152 nan 0.2306 -0.0993
## 1860 26.6105 nan 0.2306 -0.1535
## 1880 26.5182 nan 0.2306 -0.1024
## 1900 26.4134 nan 0.2306 -0.0874
## 1920 26.3600 nan 0.2306 -0.1036
## 1940 26.2909 nan 0.2306 -0.1237
## 1960 26.2837 nan 0.2306 -0.0800
## 1980 26.1490 nan 0.2306 -0.0668
## 2000 26.1309 nan 0.2306 -0.1746
## 2020 26.0259 nan 0.2306 -0.1254
## 2040 25.9625 nan 0.2306 -0.1877
## 2060 25.8981 nan 0.2306 -0.0924
## 2080 25.8165 nan 0.2306 -0.0597
## 2100 25.7430 nan 0.2306 -0.1768
## 2120 25.7011 nan 0.2306 -0.1797
## 2140 25.6292 nan 0.2306 -0.1913
## 2160 25.5840 nan 0.2306 -0.0918
## 2180 25.4846 nan 0.2306 -0.0780
## 2200 25.4759 nan 0.2306 -0.1170
## 2220 25.3391 nan 0.2306 -0.1687
## 2240 25.2853 nan 0.2306 -0.0630
## 2260 25.2516 nan 0.2306 -0.1415
## 2280 25.1985 nan 0.2306 -0.0971
## 2300 25.1492 nan 0.2306 -0.0996
## 2320 25.1080 nan 0.2306 -0.0884
## 2340 25.0750 nan 0.2306 -0.2109
## 2360 25.0159 nan 0.2306 -0.0539
## 2380 24.9945 nan 0.2306 -0.0888
## 2400 24.9641 nan 0.2306 -0.0985
## 2420 24.9471 nan 0.2306 -0.0840
## 2440 24.9258 nan 0.2306 -0.1380
## 2460 24.9521 nan 0.2306 -0.2201
## 2480 24.8717 nan 0.2306 -0.1195
## 2500 24.7740 nan 0.2306 -0.1952
## 2520 24.7201 nan 0.2306 -0.1001
## 2540 24.6765 nan 0.2306 -0.0915
## 2560 24.5981 nan 0.2306 -0.1372
## 2580 24.5648 nan 0.2306 -0.0678
## 2600 24.5252 nan 0.2306 -0.1156
## 2620 24.5237 nan 0.2306 -0.2226
## 2640 24.4851 nan 0.2306 -0.1616
## 2660 24.4549 nan 0.2306 -0.1026
## 2680 24.4833 nan 0.2306 -0.1414
## 2700 24.4485 nan 0.2306 -0.0882
## 2720 24.3531 nan 0.2306 -0.1148
## 2740 24.3235 nan 0.2306 -0.1350
## 2760 24.2847 nan 0.2306 -0.1223
## 2780 24.2358 nan 0.2306 -0.0864
## 2800 24.2400 nan 0.2306 -0.1381
## 2820 24.2132 nan 0.2306 -0.1617
## 2840 24.1837 nan 0.2306 -0.0807
## 2860 24.1258 nan 0.2306 -0.1203
## 2880 24.0677 nan 0.2306 -0.1647
## 2900 24.0136 nan 0.2306 -0.1749
## 2920 24.0322 nan 0.2306 -0.1031
## 2940 23.9735 nan 0.2306 -0.0967
## 2960 23.9715 nan 0.2306 -0.1353
## 2980 23.9069 nan 0.2306 -0.1320
## 3000 23.8990 nan 0.2306 -0.1663
## 3020 23.8338 nan 0.2306 -0.1344
## 3040 23.8472 nan 0.2306 -0.1727
## 3060 23.8241 nan 0.2306 -0.1527
## 3080 23.7739 nan 0.2306 -0.1694
## 3100 23.7631 nan 0.2306 -0.1575
## 3120 23.7235 nan 0.2306 -0.1372
## 3140 23.6874 nan 0.2306 -0.1348
## 3160 23.6848 nan 0.2306 -0.1581
## 3180 23.6543 nan 0.2306 -0.1948
## 3200 23.6799 nan 0.2306 -0.0626
## 3220 23.5944 nan 0.2306 -0.1155
## 3240 23.5806 nan 0.2306 -0.1164
## 3260 23.5217 nan 0.2306 -0.0652
## 3280 23.5373 nan 0.2306 -0.1292
## 3300 23.4803 nan 0.2306 -0.0893
## 3320 23.4941 nan 0.2306 -0.0674
## 3340 23.5074 nan 0.2306 -0.1442
## 3360 23.4615 nan 0.2306 -0.0658
## 3380 23.4297 nan 0.2306 -0.1027
## 3400 23.4043 nan 0.2306 -0.1562
## 3420 23.4015 nan 0.2306 -0.1173
## 3440 23.3682 nan 0.2306 -0.1849
## 3460 23.3731 nan 0.2306 -0.1073
## 3480 23.3273 nan 0.2306 -0.1279
## 3500 23.3003 nan 0.2306 -0.0772
## 3520 23.2587 nan 0.2306 -0.1240
## 3540 23.2272 nan 0.2306 -0.0742
## 3560 23.1933 nan 0.2306 -0.0610
## 3580 23.1425 nan 0.2306 -0.1024
## 3600 23.1123 nan 0.2306 -0.1479
## 3620 23.0701 nan 0.2306 -0.1412
## 3640 23.0412 nan 0.2306 -0.1830
## 3660 23.1056 nan 0.2306 -0.0612
## 3680 23.0661 nan 0.2306 -0.0723
## 3700 23.0953 nan 0.2306 -0.2248
## 3720 23.0461 nan 0.2306 -0.1544
##
## Iter TrainDeviance ValidDeviance StepSize Improve
## 1 678.7697 nan 0.3896 503.1787
## 2 473.7464 nan 0.3896 204.6289
## 3 374.5152 nan 0.3896 96.9691
## 4 325.4351 nan 0.3896 47.2676
## 5 296.8331 nan 0.3896 29.2309
## 6 282.8396 nan 0.3896 13.2254
## 7 273.5464 nan 0.3896 8.1405
## 8 265.6337 nan 0.3896 6.7206
## 9 258.9429 nan 0.3896 5.5590
## 10 249.2657 nan 0.3896 8.6800
## 20 210.5262 nan 0.3896 1.0040
## 40 174.4130 nan 0.3896 0.4974
## 60 157.8671 nan 0.3896 0.2055
## 80 145.2970 nan 0.3896 -0.2202
## 100 135.6379 nan 0.3896 0.3472
## 120 127.5592 nan 0.3896 -0.1042
## 140 121.9216 nan 0.3896 -0.1611
## 160 116.9891 nan 0.3896 -0.0657
## 180 112.7789 nan 0.3896 -0.1919
## 200 109.5113 nan 0.3896 -0.2648
## 220 106.2491 nan 0.3896 -0.1419
## 240 102.7174 nan 0.3896 -0.1753
## 260 100.3076 nan 0.3896 -0.2457
## 280 98.1915 nan 0.3896 -0.1277
## 300 95.9194 nan 0.3896 -0.1300
## 320 93.5540 nan 0.3896 -0.0917
## 340 91.5172 nan 0.3896 -0.1340
## 360 89.4792 nan 0.3896 -0.1869
## 380 87.6528 nan 0.3896 -0.1588
## 400 86.1073 nan 0.3896 -0.1208
## 420 84.6157 nan 0.3896 -0.3042
## 440 83.1500 nan 0.3896 -0.2447
## 460 81.4441 nan 0.3896 -0.1307
## 480 80.2811 nan 0.3896 -0.1833
## 500 78.7158 nan 0.3896 -0.1092
## 520 77.8450 nan 0.3896 -0.0438
## 540 76.5861 nan 0.3896 -0.1696
## 560 75.3056 nan 0.3896 -0.2612
## 580 73.9105 nan 0.3896 -0.1296
## 600 73.1520 nan 0.3896 -0.2570
## 620 72.2562 nan 0.3896 -0.1377
## 640 71.3256 nan 0.3896 -0.1716
## 660 70.3970 nan 0.3896 -0.1325
## 680 69.5589 nan 0.3896 -0.1638
## 700 68.6190 nan 0.3896 -0.0585
## 720 67.8047 nan 0.3896 -0.1175
## 740 66.9259 nan 0.3896 -0.2552
## 760 66.2430 nan 0.3896 -0.1314
## 780 65.4954 nan 0.3896 -0.0568
## 800 64.6903 nan 0.3896 -0.1308
## 820 63.9675 nan 0.3896 -0.1172
## 840 63.3513 nan 0.3896 -0.2136
## 860 62.8535 nan 0.3896 -0.1228
## 880 62.2590 nan 0.3896 -0.1148
## 900 61.5711 nan 0.3896 -0.0429
## 920 61.0919 nan 0.3896 -0.1043
## 940 60.6473 nan 0.3896 -0.0869
## 960 60.0386 nan 0.3896 -0.0484
## 980 59.4946 nan 0.3896 -0.1240
## 1000 58.8383 nan 0.3896 -0.1157
## 1020 58.3256 nan 0.3896 -0.0937
## 1040 57.7263 nan 0.3896 -0.1029
## 1060 57.3149 nan 0.3896 -0.0688
## 1080 57.0129 nan 0.3896 -0.1248
## 1100 56.7178 nan 0.3896 -0.0949
## 1120 56.1230 nan 0.3896 -0.1216
## 1140 55.6542 nan 0.3896 -0.0873
## 1160 55.2210 nan 0.3896 -0.1439
## 1180 54.7752 nan 0.3896 -0.1886
## 1200 54.2996 nan 0.3896 -0.1128
## 1220 53.8903 nan 0.3896 -0.1219
## 1240 53.5837 nan 0.3896 -0.1591
## 1260 53.2952 nan 0.3896 -0.1530
## 1280 53.0067 nan 0.3896 -0.0781
## 1300 52.6151 nan 0.3896 -0.1468
## 1320 52.2411 nan 0.3896 -0.1849
## 1340 51.7698 nan 0.3896 -0.2023
## 1360 51.4921 nan 0.3896 -0.0616
## 1380 51.1168 nan 0.3896 -0.1087
## 1400 50.8279 nan 0.3896 -0.1054
## 1420 50.5784 nan 0.3896 -0.1712
## 1440 50.3571 nan 0.3896 -0.2900
## 1460 50.0429 nan 0.3896 -0.0670
## 1480 49.7357 nan 0.3896 -0.1057
## 1500 49.4614 nan 0.3896 -0.0632
## 1520 49.0396 nan 0.3896 -0.0912
## 1540 48.8563 nan 0.3896 -0.0754
## 1560 48.6417 nan 0.3896 -0.1915
## 1580 48.4499 nan 0.3896 -0.0879
## 1600 48.1809 nan 0.3896 -0.0871
## 1620 47.8943 nan 0.3896 -0.0379
## 1640 47.6609 nan 0.3896 -0.0941
## 1660 47.5230 nan 0.3896 -0.0630
## 1680 47.1871 nan 0.3896 -0.1621
## 1700 46.9943 nan 0.3896 -0.1446
## 1720 46.7638 nan 0.3896 -0.0579
## 1740 46.5576 nan 0.3896 -0.2008
## 1760 46.4219 nan 0.3896 -0.0927
## 1780 46.2024 nan 0.3896 -0.0917
## 1800 45.9276 nan 0.3896 -0.0284
## 1820 45.6417 nan 0.3896 -0.1532
## 1840 45.4427 nan 0.3896 -0.0667
## 1860 45.2601 nan 0.3896 -0.0989
## 1880 45.0145 nan 0.3896 -0.1633
## 1900 44.9091 nan 0.3896 -0.1027
## 1920 44.5459 nan 0.3896 -0.1513
## 1940 44.3379 nan 0.3896 -0.0445
## 1960 44.1742 nan 0.3896 -0.1214
## 1980 43.9255 nan 0.3896 -0.0640
## 2000 43.7273 nan 0.3896 -0.1098
## 2020 43.5350 nan 0.3896 -0.1110
## 2040 43.3342 nan 0.3896 -0.1293
## 2060 43.0723 nan 0.3896 -0.1037
## 2080 42.8954 nan 0.3896 -0.1281
## 2100 42.7022 nan 0.3896 -0.0712
## 2120 42.4837 nan 0.3896 -0.1386
## 2140 42.3553 nan 0.3896 -0.1329
## 2160 42.1339 nan 0.3896 -0.1351
## 2180 42.0330 nan 0.3896 -0.1058
## 2200 41.8915 nan 0.3896 -0.1018
## 2220 41.7346 nan 0.3896 -0.1206
## 2240 41.5618 nan 0.3896 -0.1436
## 2260 41.3557 nan 0.3896 -0.1109
## 2280 41.1918 nan 0.3896 -0.0482
## 2300 41.0792 nan 0.3896 -0.1395
## 2320 40.8760 nan 0.3896 -0.0376
## 2340 40.7965 nan 0.3896 -0.0965
## 2360 40.5925 nan 0.3896 -0.0687
## 2380 40.4811 nan 0.3896 -0.1498
## 2400 40.3469 nan 0.3896 -0.1134
## 2420 40.2298 nan 0.3896 -0.1467
## 2440 40.0854 nan 0.3896 -0.5026
## 2460 39.9772 nan 0.3896 -0.1457
## 2480 39.7427 nan 0.3896 -0.1953
## 2500 39.6248 nan 0.3896 -0.1268
## 2520 39.5458 nan 0.3896 -0.1192
## 2540 39.4648 nan 0.3896 -0.0748
## 2560 39.4122 nan 0.3896 -0.1052
## 2580 39.2953 nan 0.3896 -0.0934
## 2600 39.1861 nan 0.3896 -0.2995
## 2620 39.1411 nan 0.3896 -0.1109
## 2640 38.9863 nan 0.3896 -0.0694
## 2660 38.8770 nan 0.3896 -0.0796
## 2680 38.7394 nan 0.3896 -0.1144
## 2700 38.6302 nan 0.3896 -0.0656
## 2720 38.5204 nan 0.3896 -0.1515
## 2740 38.4068 nan 0.3896 -0.0789
## 2760 38.3130 nan 0.3896 -0.1821
## 2780 38.2335 nan 0.3896 -0.0887
## 2800 38.1104 nan 0.3896 -0.0881
## 2820 38.0686 nan 0.3896 -0.2139
## 2840 37.9391 nan 0.3896 -0.0260
## 2860 37.8168 nan 0.3896 -0.0968
## 2880 37.7827 nan 0.3896 -0.1478
## 2900 37.6899 nan 0.3896 -0.0366
## 2920 37.6350 nan 0.3896 -0.1538
## 2940 37.5536 nan 0.3896 -0.1585
## 2960 37.4543 nan 0.3896 -0.0876
## 2980 37.3442 nan 0.3896 -0.0879
## 3000 37.2774 nan 0.3896 -0.0968
## 3020 37.1292 nan 0.3896 -0.0695
## 3040 37.0499 nan 0.3896 -0.2096
## 3060 36.9116 nan 0.3896 -0.0915
## 3080 36.7656 nan 0.3896 -0.1112
## 3100 36.7063 nan 0.3896 -0.0415
## 3120 36.6446 nan 0.3896 -0.0586
## 3140 36.6195 nan 0.3896 -0.0913
## 3160 36.5575 nan 0.3896 -0.0911
## 3180 36.4092 nan 0.3896 -0.1764
## 3200 36.3781 nan 0.3896 -0.1029
## 3220 36.2858 nan 0.3896 -0.1165
## 3240 36.3171 nan 0.3896 -0.2225
## 3260 36.1502 nan 0.3896 -0.0785
## 3280 36.0805 nan 0.3896 -0.0377
## 3300 36.0263 nan 0.3896 -0.1591
## 3320 36.0143 nan 0.3896 -0.1038
## 3340 35.9272 nan 0.3896 -0.1381
## 3360 35.8829 nan 0.3896 -0.1128
## 3380 35.7968 nan 0.3896 -0.1384
## 3400 35.7431 nan 0.3896 -0.0686
## 3420 35.5831 nan 0.3896 -0.0942
## 3440 35.5588 nan 0.3896 -0.1479
## 3460 35.4656 nan 0.3896 -0.1668
## 3480 35.4356 nan 0.3896 -0.0920
## 3500 35.4153 nan 0.3896 -0.1039
## 3520 35.3517 nan 0.3896 -0.1287
## 3540 35.2191 nan 0.3896 -0.0881
## 3560 35.1673 nan 0.3896 -0.0672
## 3580 35.1571 nan 0.3896 -0.0747
## 3600 35.0397 nan 0.3896 -0.1368
## 3620 35.0185 nan 0.3896 -0.0379
## 3640 34.9427 nan 0.3896 -0.0541
## 3660 34.8696 nan 0.3896 -0.1064
## 3680 34.8058 nan 0.3896 -0.1344
## 3700 34.7319 nan 0.3896 -0.1003
## 3720 34.6140 nan 0.3896 -0.1123
## 3740 34.6236 nan 0.3896 -0.1274
## 3760 34.5303 nan 0.3896 -0.1421
## 3780 34.4588 nan 0.3896 -0.1580
## 3800 34.4329 nan 0.3896 -0.0543
## 3820 34.3954 nan 0.3896 -0.0977
## 3840 34.3589 nan 0.3896 -0.0535
## 3860 34.1564 nan 0.3896 -0.0865
## 3880 34.1175 nan 0.3896 -0.0482
## 3900 34.0697 nan 0.3896 -0.0601
## 3920 33.9709 nan 0.3896 -0.1098
## 3940 33.8426 nan 0.3896 -0.1177
## 3960 33.8286 nan 0.3896 -0.0703
## 3980 33.7368 nan 0.3896 -0.0579
## 4000 33.6601 nan 0.3896 -0.0548
## 4020 33.6051 nan 0.3896 -0.0419
## 4040 33.5224 nan 0.3896 -0.0898
## 4060 33.4392 nan 0.3896 -0.0639
## 4080 33.4156 nan 0.3896 -0.0326
##
## Iter TrainDeviance ValidDeviance StepSize Improve
## 1 539.3717 nan 0.5470 631.2045
## 2 376.1753 nan 0.5470 160.7672
## 3 315.0293 nan 0.5470 63.0649
## 4 288.4488 nan 0.5470 25.0061
## 5 269.4752 nan 0.5470 17.9580
## 6 261.9994 nan 0.5470 5.5769
## 7 251.0887 nan 0.5470 10.0484
## 8 244.1084 nan 0.5470 5.6937
## 9 236.7480 nan 0.5470 6.5185
## 10 230.6240 nan 0.5470 5.0779
## 20 194.3004 nan 0.5470 1.9813
## 40 163.4400 nan 0.5470 -0.5351
## 60 145.6628 nan 0.5470 0.4153
## 80 134.6843 nan 0.5470 0.1463
## 100 125.7745 nan 0.5470 -0.4416
## 120 119.8378 nan 0.5470 0.2066
## 140 113.8232 nan 0.5470 0.0744
## 160 110.1950 nan 0.5470 -0.6106
## 180 105.6077 nan 0.5470 -0.3508
## 200 102.3102 nan 0.5470 -0.3339
## 220 98.9151 nan 0.5470 -0.1948
## 240 95.9934 nan 0.5470 -0.2652
## 260 93.4396 nan 0.5470 -0.3648
## 280 91.1500 nan 0.5470 -0.5648
## 300 88.3665 nan 0.5470 -0.0977
## 320 86.2695 nan 0.5470 -0.2783
## 340 84.1983 nan 0.5470 -0.1763
## 360 82.7655 nan 0.5470 -0.3032
## 380 81.5338 nan 0.5470 -0.2236
## 400 80.0668 nan 0.5470 -0.3131
## 420 78.7997 nan 0.5470 -0.2772
## 440 77.8022 nan 0.5470 -0.3220
## 460 76.8174 nan 0.5470 -0.2371
## 480 75.3441 nan 0.5470 -0.0937
## 500 73.7480 nan 0.5470 -0.2069
## 520 72.2020 nan 0.5470 -0.1600
## 540 71.1509 nan 0.5470 -0.2522
## 560 69.9675 nan 0.5470 -0.2386
## 580 68.9143 nan 0.5470 -0.1792
## 600 67.9822 nan 0.5470 -0.0568
## 620 67.0970 nan 0.5470 -0.1015
## 640 66.2934 nan 0.5470 -0.1129
## 660 65.5200 nan 0.5470 -0.1376
## 680 64.3896 nan 0.5470 -0.1570
## 700 63.7005 nan 0.5470 -0.1308
## 720 62.8888 nan 0.5470 -0.4394
## 740 62.1984 nan 0.5470 -0.1997
## 760 61.2829 nan 0.5470 -0.3439
## 780 60.8896 nan 0.5470 -0.1560
## 800 60.3855 nan 0.5470 -0.0890
## 820 59.8354 nan 0.5470 -0.0976
## 840 59.2943 nan 0.5470 -0.1963
## 860 58.7244 nan 0.5470 -0.2676
## 880 58.1846 nan 0.5470 -0.3888
## 900 57.5478 nan 0.5470 -0.1353
## 920 57.3608 nan 0.5470 -0.3188
## 940 56.7997 nan 0.5470 -0.2773
## 960 56.5155 nan 0.5470 -0.2262
## 980 55.7275 nan 0.5470 -0.1064
## 1000 55.1192 nan 0.5470 -0.1389
## 1020 54.6636 nan 0.5470 -0.1855
## 1040 54.4902 nan 0.5470 -0.3892
## 1060 53.8056 nan 0.5470 -0.2056
## 1080 53.3693 nan 0.5470 -0.2398
## 1100 53.1831 nan 0.5470 -0.3139
## 1120 52.6977 nan 0.5470 -0.0342
## 1140 52.3858 nan 0.5470 -0.1137
## 1160 52.0616 nan 0.5470 -0.2084
## 1180 51.6934 nan 0.5470 -0.1681
## 1200 51.2339 nan 0.5470 -0.1660
## 1220 50.7999 nan 0.5470 -0.2539
## 1240 50.4964 nan 0.5470 -0.1433
## 1260 50.2016 nan 0.5470 -0.2425
## 1280 49.8479 nan 0.5470 -0.0641
## 1300 49.5744 nan 0.5470 -0.1779
## 1320 49.1965 nan 0.5470 -0.1465
## 1340 49.0167 nan 0.5470 -0.0837
## 1360 48.6209 nan 0.5470 -0.1192
## 1380 48.3427 nan 0.5470 -0.1238
## 1400 48.0050 nan 0.5470 -0.1750
## 1420 47.8191 nan 0.5470 -0.0830
## 1440 47.5355 nan 0.5470 -0.1014
## 1460 47.2261 nan 0.5470 -0.1554
## 1480 46.9824 nan 0.5470 -0.1652
## 1500 46.6864 nan 0.5470 -0.1267
## 1520 46.5472 nan 0.5470 -0.1375
## 1540 46.3337 nan 0.5470 -0.0594
## 1560 46.0624 nan 0.5470 -0.2621
## 1580 45.8643 nan 0.5470 -0.1908
## 1600 45.5966 nan 0.5470 -0.1320
## 1620 45.2768 nan 0.5470 -0.1757
## 1640 45.0164 nan 0.5470 -0.1344
## 1660 44.8505 nan 0.5470 -0.2830
## 1680 44.7594 nan 0.5470 -0.1252
## 1700 44.4879 nan 0.5470 -0.1448
## 1720 44.3574 nan 0.5470 -0.2027
## 1740 44.1611 nan 0.5470 -0.4781
## 1760 43.9728 nan 0.5470 -0.1156
## 1780 43.7677 nan 0.5470 -0.1337
## 1800 43.6085 nan 0.5470 -0.0900
## 1820 43.5138 nan 0.5470 -0.2330
## 1840 43.2687 nan 0.5470 -0.1189
## 1860 43.0960 nan 0.5470 -0.1741
## 1880 42.8766 nan 0.5470 -0.0694
## 1900 42.8654 nan 0.5470 -0.3317
## 1920 42.6844 nan 0.5470 -0.1579
## 1940 42.6166 nan 0.5470 -0.2553
## 1960 42.4922 nan 0.5470 -0.1928
## 1980 42.2491 nan 0.5470 -0.1232
## 2000 42.2270 nan 0.5470 -0.3343
## 2020 42.1839 nan 0.5470 -0.2488
## 2040 41.9087 nan 0.5470 -0.0914
## 2060 41.8098 nan 0.5470 -0.1732
## 2080 41.5401 nan 0.5470 -0.1326
## 2100 41.4335 nan 0.5470 -0.2241
## 2120 41.3115 nan 0.5470 -0.1848
## 2140 41.0302 nan 0.5470 -0.2476
## 2160 40.8373 nan 0.5470 -0.1795
## 2180 40.8055 nan 0.5470 -0.0908
## 2200 40.5014 nan 0.5470 -0.0913
## 2220 40.3324 nan 0.5470 -0.0610
## 2240 40.3085 nan 0.5470 -0.2031
## 2260 40.1569 nan 0.5470 -0.1532
## 2280 40.1237 nan 0.5470 -0.0690
## 2300 39.8960 nan 0.5470 -0.2168
## 2320 39.7334 nan 0.5470 -0.2628
## 2340 39.5415 nan 0.5470 -0.1281
## 2360 39.4126 nan 0.5470 -0.0822
## 2380 39.3370 nan 0.5470 -0.0972
## 2400 39.2413 nan 0.5470 -0.1166
## 2420 39.1292 nan 0.5470 -0.1067
## 2440 38.8574 nan 0.5470 -0.1275
## 2460 38.9329 nan 0.5470 -0.1321
## 2480 38.7404 nan 0.5470 -0.2108
## 2500 38.6735 nan 0.5470 -0.1856
## 2520 38.5802 nan 0.5470 -0.0910
## 2540 38.4766 nan 0.5470 -0.1158
## 2560 38.3017 nan 0.5470 -0.0786
## 2580 38.1582 nan 0.5470 -0.1294
## 2600 38.0466 nan 0.5470 -0.1932
## 2620 37.9306 nan 0.5470 -0.1703
## 2640 37.8771 nan 0.5470 -0.1866
## 2660 37.7699 nan 0.5470 -0.0938
## 2680 37.6040 nan 0.5470 -0.0619
## 2700 37.4931 nan 0.5470 -0.0845
## 2720 37.4839 nan 0.5470 -0.1321
## 2740 37.4054 nan 0.5470 -0.2117
## 2760 37.3062 nan 0.5470 -0.1956
## 2780 37.2358 nan 0.5470 -0.0513
## 2800 37.1819 nan 0.5470 -0.1570
## 2820 37.0921 nan 0.5470 -0.1781
## 2840 37.0707 nan 0.5470 -0.0735
## 2860 36.9389 nan 0.5470 -0.1656
## 2880 36.9391 nan 0.5470 -0.3601
## 2900 36.7907 nan 0.5470 -0.0483
## 2920 36.6721 nan 0.5470 -0.2078
## 2940 36.6139 nan 0.5470 -0.1680
## 2960 36.5902 nan 0.5470 -0.1129
## 2980 36.4781 nan 0.5470 -0.0999
## 3000 36.4549 nan 0.5470 -0.2363
## 3020 36.3832 nan 0.5470 -0.1402
## 3040 36.2836 nan 0.5470 -0.1785
## 3060 36.3109 nan 0.5470 -0.1586
## 3080 36.2759 nan 0.5470 -0.1022
## 3100 36.1986 nan 0.5470 -0.1935
## 3120 36.0797 nan 0.5470 -0.0494
## 3140 35.9071 nan 0.5470 -0.0376
## 3160 35.8620 nan 0.5470 -0.2380
## 3180 35.8125 nan 0.5470 -0.1707
## 3200 35.7069 nan 0.5470 -0.1614
## 3220 35.6845 nan 0.5470 -0.0889
## 3240 35.5877 nan 0.5470 -0.0352
## 3260 35.5528 nan 0.5470 -0.1009
## 3280 35.4826 nan 0.5470 -0.1602
## 3300 35.4631 nan 0.5470 -0.1250
## 3320 35.4777 nan 0.5470 -0.1728
## 3340 35.3689 nan 0.5470 -0.1461
## 3360 35.2346 nan 0.5470 -0.1459
## 3380 35.2076 nan 0.5470 -0.2063
## 3400 35.2724 nan 0.5470 -0.4826
## 3420 35.0515 nan 0.5470 -0.1591
## 3440 35.0133 nan 0.5470 -0.1146
## 3460 34.9285 nan 0.5470 -0.0936
## 3480 34.8377 nan 0.5470 -0.1669
## 3489 34.7579 nan 0.5470 -0.1585
##
## Iter TrainDeviance ValidDeviance StepSize Improve
## 1 819.8765 nan 0.2306 361.8322
## 2 598.2295 nan 0.2306 224.4198
## 3 464.0540 nan 0.2306 134.2621
## 4 376.0027 nan 0.2306 86.9478
## 5 317.5825 nan 0.2306 58.9315
## 6 278.5187 nan 0.2306 38.3842
## 7 251.2128 nan 0.2306 25.6245
## 8 230.2331 nan 0.2306 20.0655
## 9 216.3586 nan 0.2306 13.3059
## 10 205.3586 nan 0.2306 9.8532
## 20 162.4408 nan 0.2306 1.4088
## 40 133.9545 nan 0.2306 0.3749
## 60 116.6463 nan 0.2306 0.4550
## 80 104.9094 nan 0.2306 0.1072
## 100 95.5531 nan 0.2306 0.2552
## 120 88.0855 nan 0.2306 0.0193
## 140 82.6312 nan 0.2306 -0.1468
## 160 77.7783 nan 0.2306 -0.0623
## 180 73.8172 nan 0.2306 -0.1295
## 200 69.9356 nan 0.2306 -0.1116
## 220 66.1016 nan 0.2306 -0.1304
## 240 63.2089 nan 0.2306 -0.0773
## 260 60.9367 nan 0.2306 -0.1313
## 280 58.5218 nan 0.2306 -0.0858
## 300 56.4094 nan 0.2306 -0.1242
## 320 54.9835 nan 0.2306 -0.2025
## 340 53.1561 nan 0.2306 -0.1028
## 360 51.3590 nan 0.2306 -0.1202
## 380 49.9335 nan 0.2306 -0.1379
## 400 48.4651 nan 0.2306 -0.2069
## 420 47.1839 nan 0.2306 -0.1518
## 440 46.0711 nan 0.2306 -0.0609
## 460 45.2544 nan 0.2306 -0.1892
## 480 44.0896 nan 0.2306 -0.1681
## 500 43.1091 nan 0.2306 -0.1762
## 520 42.1875 nan 0.2306 -0.1269
## 540 41.3086 nan 0.2306 -0.0840
## 560 40.5454 nan 0.2306 -0.1971
## 580 39.9184 nan 0.2306 -0.1562
## 600 39.2210 nan 0.2306 -0.0071
## 620 38.5929 nan 0.2306 -0.1132
## 640 38.0082 nan 0.2306 -0.1333
## 660 37.4158 nan 0.2306 -0.0580
## 680 36.7591 nan 0.2306 -0.1318
## 700 36.1985 nan 0.2306 -0.0532
## 720 35.7535 nan 0.2306 -0.1721
## 740 35.1359 nan 0.2306 -0.1954
## 760 34.7109 nan 0.2306 -0.0614
## 780 34.2381 nan 0.2306 -0.1800
## 800 33.8372 nan 0.2306 -0.1747
## 820 33.4408 nan 0.2306 -0.1312
## 840 33.0957 nan 0.2306 -0.0842
## 860 32.7284 nan 0.2306 -0.0691
## 880 32.3847 nan 0.2306 -0.1260
## 900 32.0484 nan 0.2306 -0.0486
## 920 31.7746 nan 0.2306 -0.1386
## 940 31.5020 nan 0.2306 -0.1088
## 960 31.1952 nan 0.2306 -0.0896
## 980 30.9626 nan 0.2306 -0.2405
## 1000 30.7206 nan 0.2306 -0.0822
## 1020 30.4890 nan 0.2306 -0.1836
## 1040 30.2428 nan 0.2306 -0.1010
## 1060 29.9962 nan 0.2306 -0.0803
## 1080 29.7939 nan 0.2306 -0.1187
## 1100 29.5431 nan 0.2306 -0.1178
## 1120 29.4116 nan 0.2306 -0.1431
## 1140 29.1493 nan 0.2306 -0.1371
## 1160 28.9660 nan 0.2306 -0.0753
## 1180 28.7811 nan 0.2306 -0.0864
## 1200 28.5807 nan 0.2306 -0.1110
## 1220 28.4457 nan 0.2306 -0.2131
## 1240 28.2837 nan 0.2306 -0.1091
## 1260 28.1180 nan 0.2306 -0.1378
## 1280 27.9570 nan 0.2306 -0.1215
## 1300 27.7725 nan 0.2306 -0.0755
## 1320 27.6364 nan 0.2306 -0.1102
## 1340 27.5067 nan 0.2306 -0.0994
## 1360 27.4065 nan 0.2306 -0.0923
## 1380 27.2481 nan 0.2306 -0.0753
## 1400 27.0937 nan 0.2306 -0.1692
## 1420 26.9570 nan 0.2306 -0.1009
## 1440 26.8464 nan 0.2306 -0.1533
## 1460 26.6846 nan 0.2306 -0.0621
## 1480 26.6199 nan 0.2306 -0.0860
## 1500 26.5390 nan 0.2306 -0.0984
## 1520 26.3866 nan 0.2306 -0.1547
## 1540 26.2930 nan 0.2306 -0.0850
## 1560 26.2448 nan 0.2306 -0.1591
## 1580 26.1397 nan 0.2306 -0.0867
## 1600 25.9919 nan 0.2306 -0.0980
## 1620 25.9026 nan 0.2306 -0.1127
## 1640 25.7959 nan 0.2306 -0.1129
## 1660 25.7062 nan 0.2306 -0.0645
## 1680 25.6171 nan 0.2306 -0.0868
## 1700 25.5246 nan 0.2306 -0.1181
## 1720 25.3573 nan 0.2306 -0.2116
## 1740 25.2694 nan 0.2306 -0.1368
## 1760 25.1733 nan 0.2306 -0.0600
## 1780 25.0890 nan 0.2306 -0.1335
## 1800 25.0151 nan 0.2306 -0.1168
## 1820 24.9183 nan 0.2306 -0.1475
## 1840 24.8529 nan 0.2306 -0.2620
## 1860 24.7174 nan 0.2306 -0.0808
## 1880 24.6393 nan 0.2306 -0.1215
## 1900 24.5762 nan 0.2306 -0.1234
## 1920 24.5490 nan 0.2306 -0.1038
## 1940 24.4565 nan 0.2306 -0.1589
## 1960 24.4158 nan 0.2306 -0.0796
## 1980 24.3471 nan 0.2306 -0.0769
## 2000 24.2973 nan 0.2306 -0.1623
## 2020 24.2531 nan 0.2306 -0.0844
## 2040 24.1486 nan 0.2306 -0.0738
## 2060 24.1190 nan 0.2306 -0.1507
## 2080 24.0557 nan 0.2306 -0.1297
## 2100 24.0151 nan 0.2306 -0.1298
## 2120 23.9636 nan 0.2306 -0.0899
## 2140 23.9176 nan 0.2306 -0.1026
## 2160 23.8149 nan 0.2306 -0.0858
## 2180 23.7758 nan 0.2306 -0.0979
## 2200 23.7217 nan 0.2306 -0.1371
## 2220 23.6310 nan 0.2306 -0.0274
## 2240 23.6149 nan 0.2306 -0.1038
## 2260 23.5792 nan 0.2306 -0.0818
## 2280 23.5272 nan 0.2306 -0.1539
## 2300 23.4733 nan 0.2306 -0.1062
## 2320 23.4645 nan 0.2306 -0.1298
## 2340 23.3995 nan 0.2306 -0.0810
## 2360 23.3915 nan 0.2306 -0.2365
## 2380 23.3108 nan 0.2306 -0.1462
## 2400 23.2791 nan 0.2306 -0.1145
## 2420 23.2119 nan 0.2306 -0.0510
## 2440 23.2116 nan 0.2306 -0.0658
## 2460 23.1930 nan 0.2306 -0.1567
## 2480 23.1572 nan 0.2306 -0.0572
## 2500 23.1357 nan 0.2306 -0.0711
## 2520 23.0606 nan 0.2306 -0.0952
## 2540 22.9933 nan 0.2306 -0.1006
## 2560 22.9424 nan 0.2306 -0.1878
## 2580 22.9302 nan 0.2306 -0.1106
## 2600 22.8581 nan 0.2306 -0.0543
## 2620 22.8276 nan 0.2306 -0.0745
## 2640 22.7904 nan 0.2306 -0.1234
## 2660 22.7435 nan 0.2306 -0.1580
## 2680 22.7036 nan 0.2306 -0.1114
## 2700 22.7336 nan 0.2306 -0.2043
## 2720 22.6553 nan 0.2306 -0.1064
## 2740 22.6317 nan 0.2306 -0.1147
## 2760 22.6120 nan 0.2306 -0.1472
## 2780 22.5526 nan 0.2306 -0.1292
## 2800 22.5401 nan 0.2306 -0.0677
## 2820 22.5412 nan 0.2306 -0.0861
## 2840 22.5066 nan 0.2306 -0.1429
## 2860 22.4471 nan 0.2306 -0.1039
## 2880 22.4174 nan 0.2306 -0.0883
## 2900 22.4014 nan 0.2306 -0.1113
## 2920 22.3817 nan 0.2306 -0.0981
## 2940 22.3421 nan 0.2306 -0.1610
## 2960 22.2988 nan 0.2306 -0.1206
## 2980 22.3233 nan 0.2306 -0.0659
## 3000 22.3354 nan 0.2306 -0.0735
## 3020 22.2963 nan 0.2306 -0.2677
## 3040 22.2277 nan 0.2306 -0.1469
## 3060 22.2160 nan 0.2306 -0.1462
## 3080 22.1785 nan 0.2306 -0.1610
## 3100 22.1599 nan 0.2306 -0.0888
## 3120 22.1013 nan 0.2306 -0.1222
## 3140 22.1065 nan 0.2306 -0.0972
## 3160 22.1006 nan 0.2306 -0.1395
## 3180 22.0977 nan 0.2306 -0.0531
## 3200 22.0941 nan 0.2306 -0.0683
## 3220 22.0509 nan 0.2306 -0.0946
## 3240 22.0357 nan 0.2306 -0.1127
## 3260 22.0067 nan 0.2306 -0.0866
## 3280 21.9931 nan 0.2306 -0.1395
## 3300 21.9409 nan 0.2306 -0.0449
## 3320 21.9419 nan 0.2306 -0.2252
## 3340 21.8617 nan 0.2306 -0.1438
## 3360 21.8288 nan 0.2306 -0.0732
## 3380 21.8243 nan 0.2306 -0.1021
## 3400 21.7973 nan 0.2306 -0.0925
## 3420 21.7410 nan 0.2306 -0.1755
## 3440 21.7310 nan 0.2306 -0.0538
## 3460 21.6749 nan 0.2306 -0.1795
## 3480 21.6333 nan 0.2306 -0.0701
## 3500 21.6090 nan 0.2306 -0.1037
## 3520 21.6528 nan 0.2306 -0.2131
## 3540 21.6204 nan 0.2306 -0.1364
## 3560 21.6011 nan 0.2306 -0.1179
## 3580 21.5847 nan 0.2306 -0.2377
## 3600 21.5815 nan 0.2306 -0.0931
## 3620 21.5341 nan 0.2306 -0.0630
## 3640 21.5170 nan 0.2306 -0.0803
## 3660 21.4780 nan 0.2306 -0.0954
## 3680 21.5298 nan 0.2306 -0.0978
## 3700 21.4791 nan 0.2306 -0.0835
## 3720 21.4705 nan 0.2306 -0.1347
##
## Iter TrainDeviance ValidDeviance StepSize Improve
## 1 673.9031 nan 0.3896 500.2957
## 2 455.4256 nan 0.3896 216.0964
## 3 358.2185 nan 0.3896 95.1424
## 4 309.8896 nan 0.3896 47.2781
## 5 283.2738 nan 0.3896 25.4522
## 6 266.1795 nan 0.3896 16.0087
## 7 250.3216 nan 0.3896 14.0969
## 8 242.5260 nan 0.3896 7.8769
## 9 236.6881 nan 0.3896 5.4792
## 10 230.4422 nan 0.3896 5.6839
## 20 194.2343 nan 0.3896 1.0120
## 40 164.7300 nan 0.3896 0.8619
## 60 148.9902 nan 0.3896 0.1639
## 80 137.4391 nan 0.3896 -0.0071
## 100 129.5421 nan 0.3896 -0.2870
## 120 121.3237 nan 0.3896 -0.1098
## 140 116.5812 nan 0.3896 -0.1783
## 160 111.5619 nan 0.3896 -0.3795
## 180 107.8271 nan 0.3896 -0.1306
## 200 104.3140 nan 0.3896 -0.3044
## 220 100.8734 nan 0.3896 -0.1528
## 240 97.6328 nan 0.3896 -0.1211
## 260 95.2757 nan 0.3896 -0.0616
## 280 92.5858 nan 0.3896 -0.1949
## 300 90.0852 nan 0.3896 -0.1168
## 320 88.2393 nan 0.3896 -0.2722
## 340 86.2874 nan 0.3896 -0.0742
## 360 84.5821 nan 0.3896 -0.1664
## 380 82.8974 nan 0.3896 -0.0724
## 400 81.1764 nan 0.3896 -0.0969
## 420 79.6765 nan 0.3896 -0.3535
## 440 78.2627 nan 0.3896 -0.1443
## 460 76.8900 nan 0.3896 -0.2163
## 480 75.7343 nan 0.3896 -0.1056
## 500 74.2103 nan 0.3896 -0.1123
## 520 73.1068 nan 0.3896 -0.1164
## 540 71.8049 nan 0.3896 -0.0963
## 560 70.8381 nan 0.3896 -0.1184
## 580 69.6085 nan 0.3896 -0.1134
## 600 68.6873 nan 0.3896 0.0102
## 620 67.6131 nan 0.3896 -0.1351
## 640 66.6227 nan 0.3896 -0.1512
## 660 65.7311 nan 0.3896 -0.2059
## 680 64.8981 nan 0.3896 -0.2007
## 700 64.1122 nan 0.3896 -0.1255
## 720 63.3810 nan 0.3896 -0.0623
## 740 62.8539 nan 0.3896 -0.1732
## 760 62.1208 nan 0.3896 -0.2411
## 780 61.4335 nan 0.3896 -0.1531
## 800 60.8025 nan 0.3896 -0.0080
## 820 60.2368 nan 0.3896 -0.0954
## 840 59.4483 nan 0.3896 -0.1027
## 860 58.6776 nan 0.3896 -0.1093
## 880 58.3058 nan 0.3896 -0.1662
## 900 57.5642 nan 0.3896 -0.0783
## 920 56.9282 nan 0.3896 -0.1450
## 940 56.5008 nan 0.3896 -0.0904
## 960 56.0618 nan 0.3896 -0.1093
## 980 55.6596 nan 0.3896 -0.1709
## 1000 55.1015 nan 0.3896 -0.0890
## 1020 54.5777 nan 0.3896 -0.1376
## 1040 54.1125 nan 0.3896 -0.1166
## 1060 53.6697 nan 0.3896 -0.2745
## 1080 53.1394 nan 0.3896 -0.1658
## 1100 52.8326 nan 0.3896 -0.2075
## 1120 52.2829 nan 0.3896 -0.1101
## 1140 51.9782 nan 0.3896 -0.0896
## 1160 51.4099 nan 0.3896 -0.0612
## 1180 50.9002 nan 0.3896 -0.1619
## 1200 50.6390 nan 0.3896 -0.0915
## 1220 50.2102 nan 0.3896 -0.1754
## 1240 49.9869 nan 0.3896 -0.1358
## 1260 49.6688 nan 0.3896 -0.1224
## 1280 49.2512 nan 0.3896 -0.1184
## 1300 49.0323 nan 0.3896 -0.0517
## 1320 48.6898 nan 0.3896 -0.0957
## 1340 48.3388 nan 0.3896 -0.1299
## 1360 48.0321 nan 0.3896 -0.1919
## 1380 47.7014 nan 0.3896 -0.1383
## 1400 47.4831 nan 0.3896 -0.1025
## 1420 47.2306 nan 0.3896 -0.1188
## 1440 47.1072 nan 0.3896 -0.0826
## 1460 46.8655 nan 0.3896 -0.1188
## 1480 46.6476 nan 0.3896 -0.1930
## 1500 46.2485 nan 0.3896 -0.0328
## 1520 45.9593 nan 0.3896 -0.0029
## 1540 45.6805 nan 0.3896 -0.0828
## 1560 45.4350 nan 0.3896 -0.1041
## 1580 45.2622 nan 0.3896 -0.3048
## 1600 45.0871 nan 0.3896 -0.1827
## 1620 44.8916 nan 0.3896 -0.0988
## 1640 44.5624 nan 0.3896 -0.0544
## 1660 44.2683 nan 0.3896 -0.0989
## 1680 43.9999 nan 0.3896 -0.1739
## 1700 43.7996 nan 0.3896 -0.1818
## 1720 43.6401 nan 0.3896 -0.0999
## 1740 43.3890 nan 0.3896 -0.0940
## 1760 43.2293 nan 0.3896 -0.1083
## 1780 43.0082 nan 0.3896 -0.1001
## 1800 42.7921 nan 0.3896 -0.1731
## 1820 42.6738 nan 0.3896 -0.1008
## 1840 42.4238 nan 0.3896 -0.1324
## 1860 42.2028 nan 0.3896 -0.1106
## 1880 42.0306 nan 0.3896 -0.1672
## 1900 41.7389 nan 0.3896 -0.1013
## 1920 41.6192 nan 0.3896 -0.1463
## 1940 41.4178 nan 0.3896 -0.0847
## 1960 41.2259 nan 0.3896 -0.0570
## 1980 41.0306 nan 0.3896 -0.1492
## 2000 40.8690 nan 0.3896 -0.1067
## 2020 40.7217 nan 0.3896 -0.0946
## 2040 40.6477 nan 0.3896 -0.2288
## 2060 40.4525 nan 0.3896 -0.1078
## 2080 40.3123 nan 0.3896 -0.1624
## 2100 40.2133 nan 0.3896 -0.1123
## 2120 40.0696 nan 0.3896 -0.1103
## 2140 39.9090 nan 0.3896 -0.1153
## 2160 39.8397 nan 0.3896 -0.1601
## 2180 39.6905 nan 0.3896 -0.0474
## 2200 39.5341 nan 0.3896 -0.0628
## 2220 39.3501 nan 0.3896 -0.0955
## 2240 39.1995 nan 0.3896 -0.0612
## 2260 39.0361 nan 0.3896 -0.0470
## 2280 38.9398 nan 0.3896 -0.1767
## 2300 38.7996 nan 0.3896 -0.1557
## 2320 38.6542 nan 0.3896 -0.0701
## 2340 38.5434 nan 0.3896 -0.0616
## 2360 38.4232 nan 0.3896 -0.1369
## 2380 38.3354 nan 0.3896 -0.1177
## 2400 38.2002 nan 0.3896 -0.0640
## 2420 38.0974 nan 0.3896 -0.1246
## 2440 37.9927 nan 0.3896 -0.0709
## 2460 37.8229 nan 0.3896 -0.0926
## 2480 37.5999 nan 0.3896 -0.0659
## 2500 37.5115 nan 0.3896 -0.1029
## 2520 37.4422 nan 0.3896 -0.0773
## 2540 37.2669 nan 0.3896 -0.0394
## 2560 37.2045 nan 0.3896 -0.0621
## 2580 37.1051 nan 0.3896 -0.1339
## 2600 37.0232 nan 0.3896 -0.0317
## 2620 36.9208 nan 0.3896 -0.0977
## 2640 36.8658 nan 0.3896 -0.1219
## 2660 36.7807 nan 0.3896 -0.1460
## 2680 36.6335 nan 0.3896 -0.0238
## 2700 36.4513 nan 0.3896 -0.0257
## 2720 36.3483 nan 0.3896 -0.0796
## 2740 36.2104 nan 0.3896 -0.0329
## 2760 36.1544 nan 0.3896 -0.0935
## 2780 36.0480 nan 0.3896 -0.1642
## 2800 35.9996 nan 0.3896 -0.0881
## 2820 35.8920 nan 0.3896 -0.1034
## 2840 35.8172 nan 0.3896 -0.0776
## 2860 35.7377 nan 0.3896 -0.0943
## 2880 35.5685 nan 0.3896 -0.0618
## 2900 35.4652 nan 0.3896 -0.0431
## 2920 35.3710 nan 0.3896 -0.0478
## 2940 35.2530 nan 0.3896 -0.0681
## 2960 35.1941 nan 0.3896 -0.0964
## 2980 35.0997 nan 0.3896 -0.0743
## 3000 35.0411 nan 0.3896 -0.0958
## 3020 34.9146 nan 0.3896 -0.1337
## 3040 34.9076 nan 0.3896 -0.0920
## 3060 34.8664 nan 0.3896 -0.1061
## 3080 34.7441 nan 0.3896 -0.0414
## 3100 34.6495 nan 0.3896 -0.0659
## 3120 34.6451 nan 0.3896 -0.2376
## 3140 34.5214 nan 0.3896 -0.1166
## 3160 34.5298 nan 0.3896 -0.2053
## 3180 34.4098 nan 0.3896 -0.0448
## 3200 34.3777 nan 0.3896 -0.1166
## 3220 34.2705 nan 0.3896 -0.0598
## 3240 34.1637 nan 0.3896 -0.0947
## 3260 34.0411 nan 0.3896 -0.0405
## 3280 33.9546 nan 0.3896 -0.0693
## 3300 33.8490 nan 0.3896 -0.0949
## 3320 33.7447 nan 0.3896 -0.0989
## 3340 33.6874 nan 0.3896 -0.0559
## 3360 33.5369 nan 0.3896 -0.0936
## 3380 33.5626 nan 0.3896 -0.0733
## 3400 33.4417 nan 0.3896 -0.0806
## 3420 33.3873 nan 0.3896 -0.0809
## 3440 33.3274 nan 0.3896 -0.0831
## 3460 33.2429 nan 0.3896 -0.1112
## 3480 33.0975 nan 0.3896 -0.0490
## 3500 32.9955 nan 0.3896 -0.0809
## 3520 32.9825 nan 0.3896 -0.0294
## 3540 32.9453 nan 0.3896 -0.1194
## 3560 32.8637 nan 0.3896 -0.0783
## 3580 32.8467 nan 0.3896 -0.1302
## 3600 32.7580 nan 0.3896 -0.0525
## 3620 32.6683 nan 0.3896 -0.0769
## 3640 32.6107 nan 0.3896 -0.0835
## 3660 32.5858 nan 0.3896 -0.1847
## 3680 32.4910 nan 0.3896 -0.0526
## 3700 32.4548 nan 0.3896 -0.1408
## 3720 32.3622 nan 0.3896 -0.0413
## 3740 32.2923 nan 0.3896 -0.0640
## 3760 32.2367 nan 0.3896 -0.0802
## 3780 32.2408 nan 0.3896 -0.1045
## 3800 32.2339 nan 0.3896 -0.0776
## 3820 32.1961 nan 0.3896 -0.0477
## 3840 32.2337 nan 0.3896 -0.2785
## 3860 32.1050 nan 0.3896 -0.0942
## 3880 32.0525 nan 0.3896 -0.1175
## 3900 31.9834 nan 0.3896 -0.1561
## 3920 31.9580 nan 0.3896 -0.1473
## 3940 31.8972 nan 0.3896 -0.0689
## 3960 31.8321 nan 0.3896 -0.0928
## 3980 31.7670 nan 0.3896 -0.0869
## 4000 31.6861 nan 0.3896 -0.0904
## 4020 31.7139 nan 0.3896 -0.1142
## 4040 31.6407 nan 0.3896 -0.0590
## 4060 31.6072 nan 0.3896 -0.1380
## 4080 31.5295 nan 0.3896 -0.1541
##
## Iter TrainDeviance ValidDeviance StepSize Improve
## 1 543.4883 nan 0.5470 640.5988
## 2 373.4748 nan 0.5470 172.0838
## 3 315.5182 nan 0.5470 57.1656
## 4 282.7272 nan 0.5470 32.5032
## 5 266.6091 nan 0.5470 14.7395
## 6 253.4118 nan 0.5470 12.0296
## 7 244.1036 nan 0.5470 8.0444
## 8 237.6290 nan 0.5470 5.7521
## 9 231.7601 nan 0.5470 4.3249
## 10 226.2360 nan 0.5470 5.4128
## 20 195.5105 nan 0.5470 0.8227
## 40 163.9015 nan 0.5470 1.2057
## 60 149.5919 nan 0.5470 -0.6698
## 80 137.1281 nan 0.5470 -0.3005
## 100 129.1597 nan 0.5470 -0.0139
## 120 121.3919 nan 0.5470 -0.4264
## 140 115.8118 nan 0.5470 -0.0071
## 160 110.1566 nan 0.5470 0.0976
## 180 104.9477 nan 0.5470 0.0538
## 200 101.2229 nan 0.5470 -0.3254
## 220 98.2817 nan 0.5470 -0.1683
## 240 94.6229 nan 0.5470 -0.2898
## 260 92.3571 nan 0.5470 -0.3402
## 280 89.8366 nan 0.5470 -0.2887
## 300 87.7458 nan 0.5470 -0.0441
## 320 85.9497 nan 0.5470 0.1917
## 340 83.8035 nan 0.5470 -0.2869
## 360 81.3411 nan 0.5470 -0.2527
## 380 79.6566 nan 0.5470 -0.3291
## 400 78.3513 nan 0.5470 -0.4047
## 420 76.9632 nan 0.5470 -0.1788
## 440 75.5978 nan 0.5470 -0.1070
## 460 74.5064 nan 0.5470 -0.1740
## 480 73.2707 nan 0.5470 -0.2416
## 500 71.9748 nan 0.5470 -0.2122
## 520 70.7853 nan 0.5470 -0.2448
## 540 69.6390 nan 0.5470 -0.0512
## 560 68.8527 nan 0.5470 -0.1742
## 580 67.5710 nan 0.5470 -0.3371
## 600 66.7825 nan 0.5470 -0.1340
## 620 65.7939 nan 0.5470 -0.0869
## 640 64.8150 nan 0.5470 -0.0667
## 660 63.6550 nan 0.5470 -0.1584
## 680 63.0013 nan 0.5470 -0.3063
## 700 62.1851 nan 0.5470 -0.1820
## 720 61.4775 nan 0.5470 -0.2282
## 740 60.8895 nan 0.5470 -0.2705
## 760 60.1433 nan 0.5470 -0.2651
## 780 59.5230 nan 0.5470 -0.2296
## 800 58.7180 nan 0.5470 -0.1087
## 820 58.2527 nan 0.5470 -0.3931
## 840 57.6474 nan 0.5470 -0.1033
## 860 56.8383 nan 0.5470 -0.0729
## 880 56.3397 nan 0.5470 -0.1092
## 900 55.7581 nan 0.5470 0.0013
## 920 55.0518 nan 0.5470 -0.1712
## 940 54.4409 nan 0.5470 -0.0672
## 960 53.8321 nan 0.5470 -0.2265
## 980 53.4614 nan 0.5470 -0.0993
## 1000 53.0070 nan 0.5470 -0.1635
## 1020 52.6584 nan 0.5470 -0.1540
## 1040 52.3592 nan 0.5470 -0.2909
## 1060 51.8507 nan 0.5470 -0.1433
## 1080 51.4615 nan 0.5470 -0.1640
## 1100 50.8890 nan 0.5470 -0.1814
## 1120 50.6278 nan 0.5470 -0.2431
## 1140 50.2023 nan 0.5470 0.0025
## 1160 49.6968 nan 0.5470 -0.0787
## 1180 49.3723 nan 0.5470 -0.4154
## 1200 49.0834 nan 0.5470 -0.1259
## 1220 48.7116 nan 0.5470 -0.1851
## 1240 48.2451 nan 0.5470 -0.1798
## 1260 47.9123 nan 0.5470 -0.0727
## 1280 47.6716 nan 0.5470 -0.2048
## 1300 47.2868 nan 0.5470 -0.1093
## 1320 46.9720 nan 0.5470 -0.2103
## 1340 46.6454 nan 0.5470 -0.0733
## 1360 46.4821 nan 0.5470 -0.1588
## 1380 46.1590 nan 0.5470 0.0193
## 1400 45.9975 nan 0.5470 -0.1564
## 1420 45.6874 nan 0.5470 -0.1777
## 1440 45.4389 nan 0.5470 -0.1775
## 1460 45.2446 nan 0.5470 -0.2059
## 1480 44.9184 nan 0.5470 0.0429
## 1500 44.7380 nan 0.5470 -0.1351
## 1520 44.5496 nan 0.5470 -0.1632
## 1540 44.3400 nan 0.5470 -0.1109
## 1560 44.1774 nan 0.5470 -0.2138
## 1580 43.9117 nan 0.5470 -0.0722
## 1600 43.7284 nan 0.5470 -0.1555
## 1620 43.5310 nan 0.5470 -0.1258
## 1640 43.2050 nan 0.5470 -0.1549
## 1660 43.0115 nan 0.5470 -0.1106
## 1680 42.6874 nan 0.5470 -0.1546
## 1700 42.5185 nan 0.5470 -0.1168
## 1720 42.4123 nan 0.5470 -0.1529
## 1740 42.1947 nan 0.5470 -0.1621
## 1760 41.8812 nan 0.5470 -0.2319
## 1780 41.7542 nan 0.5470 -0.1417
## 1800 41.7704 nan 0.5470 -0.2929
## 1820 41.5211 nan 0.5470 -0.1072
## 1840 41.4024 nan 0.5470 -0.0981
## 1860 41.2777 nan 0.5470 -0.1112
## 1880 41.0254 nan 0.5470 -0.1512
## 1900 40.9554 nan 0.5470 -0.1331
## 1920 40.8087 nan 0.5470 -0.0371
## 1940 40.5816 nan 0.5470 -0.1994
## 1960 40.5270 nan 0.5470 -0.2801
## 1980 40.2339 nan 0.5470 -0.2085
## 2000 40.2109 nan 0.5470 -0.2033
## 2020 40.0113 nan 0.5470 -0.1482
## 2040 39.9071 nan 0.5470 -0.1813
## 2060 39.7081 nan 0.5470 -0.2135
## 2080 39.5292 nan 0.5470 -0.0738
## 2100 39.4006 nan 0.5470 -0.1676
## 2120 39.3752 nan 0.5470 -0.1183
## 2140 39.2138 nan 0.5470 -0.1998
## 2160 39.0565 nan 0.5470 -0.2488
## 2180 38.9535 nan 0.5470 -0.1670
## 2200 38.7862 nan 0.5470 -0.1419
## 2220 38.7810 nan 0.5470 -0.0892
## 2240 38.6108 nan 0.5470 -0.1938
## 2260 38.4350 nan 0.5470 -0.1385
## 2280 38.3410 nan 0.5470 -0.2381
## 2300 38.1556 nan 0.5470 -0.2041
## 2320 38.1520 nan 0.5470 -0.5389
## 2340 37.9377 nan 0.5470 -0.0636
## 2360 37.9202 nan 0.5470 -0.3013
## 2380 37.7444 nan 0.5470 -0.1175
## 2400 37.6146 nan 0.5470 -0.1423
## 2420 37.5770 nan 0.5470 -0.3117
## 2440 37.4795 nan 0.5470 -0.2243
## 2460 37.2826 nan 0.5470 -0.1735
## 2480 37.3931 nan 0.5470 -0.1606
## 2500 37.2550 nan 0.5470 -0.1283
## 2520 37.1331 nan 0.5470 -0.0610
## 2540 37.0108 nan 0.5470 -0.0873
## 2560 36.9261 nan 0.5470 -0.0555
## 2580 36.8990 nan 0.5470 -0.1864
## 2600 36.7652 nan 0.5470 -0.1779
## 2620 36.5827 nan 0.5470 -0.1643
## 2640 36.3740 nan 0.5470 -0.1601
## 2660 36.1991 nan 0.5470 -0.0011
## 2680 36.2092 nan 0.5470 -0.1509
## 2700 36.1222 nan 0.5470 -0.2528
## 2720 35.9968 nan 0.5470 -0.1764
## 2740 35.9361 nan 0.5470 -0.3287
## 2760 35.8817 nan 0.5470 -0.1684
## 2780 35.7672 nan 0.5470 -0.2022
## 2800 35.5962 nan 0.5470 -0.0684
## 2820 35.5327 nan 0.5470 -0.1635
## 2840 35.3906 nan 0.5470 -0.1656
## 2860 35.2345 nan 0.5470 -0.0917
## 2880 35.1213 nan 0.5470 -0.2014
## 2900 35.0848 nan 0.5470 -0.1745
## 2920 35.0482 nan 0.5470 -0.1855
## 2940 34.8702 nan 0.5470 -0.0538
## 2960 34.9113 nan 0.5470 -0.1491
## 2980 34.9093 nan 0.5470 -0.1049
## 3000 34.7691 nan 0.5470 -0.1936
## 3020 34.6902 nan 0.5470 -0.2690
## 3040 34.6935 nan 0.5470 -0.2933
## 3060 34.5735 nan 0.5470 -0.0343
## 3080 34.5341 nan 0.5470 -0.0850
## 3100 34.4176 nan 0.5470 -0.1013
## 3120 34.3971 nan 0.5470 -0.1389
## 3140 34.3392 nan 0.5470 -0.1242
## 3160 34.2075 nan 0.5470 -0.1675
## 3180 34.0633 nan 0.5470 -0.1467
## 3200 34.0964 nan 0.5470 -0.0459
## 3220 33.9575 nan 0.5470 -0.1594
## 3240 33.8873 nan 0.5470 -0.0880
## 3260 33.8063 nan 0.5470 -0.1030
## 3280 33.7274 nan 0.5470 -0.1162
## 3300 33.6564 nan 0.5470 -0.1451
## 3320 33.6013 nan 0.5470 -0.1953
## 3340 33.5415 nan 0.5470 -0.2858
## 3360 33.5413 nan 0.5470 -0.2274
## 3380 33.3867 nan 0.5470 -0.1829
## 3400 33.2448 nan 0.5470 -0.1552
## 3420 33.1647 nan 0.5470 -0.0840
## 3440 33.1560 nan 0.5470 -0.0778
## 3460 33.0858 nan 0.5470 -0.1247
## 3480 33.1523 nan 0.5470 -0.3534
## 3489 33.1368 nan 0.5470 -0.0527
##
## Iter TrainDeviance ValidDeviance StepSize Improve
## 1 825.8791 nan 0.2306 352.7870
## 2 604.6337 nan 0.2306 218.9449
## 3 464.4869 nan 0.2306 136.9837
## 4 374.0772 nan 0.2306 90.3507
## 5 313.5929 nan 0.2306 59.2327
## 6 275.5003 nan 0.2306 36.9337
## 7 250.4041 nan 0.2306 24.3839
## 8 230.1951 nan 0.2306 18.6278
## 9 218.2181 nan 0.2306 10.8484
## 10 207.4607 nan 0.2306 9.6362
## 20 166.0889 nan 0.2306 1.0924
## 40 135.8218 nan 0.2306 0.7384
## 60 118.3438 nan 0.2306 0.3476
## 80 107.1410 nan 0.2306 0.2327
## 100 98.4464 nan 0.2306 0.0169
## 120 92.5357 nan 0.2306 -0.2377
## 140 86.5172 nan 0.2306 -0.0122
## 160 81.0686 nan 0.2306 -0.0907
## 180 76.6289 nan 0.2306 0.0360
## 200 72.6668 nan 0.2306 -0.3552
## 220 69.0827 nan 0.2306 -0.1137
## 240 66.4107 nan 0.2306 -0.2068
## 260 64.3094 nan 0.2306 -0.1655
## 280 62.0427 nan 0.2306 -0.0949
## 300 59.8466 nan 0.2306 -0.1726
## 320 58.0098 nan 0.2306 -0.0726
## 340 56.3326 nan 0.2306 -0.1098
## 360 54.7531 nan 0.2306 -0.0700
## 380 53.4447 nan 0.2306 -0.1428
## 400 52.1125 nan 0.2306 -0.0991
## 420 50.7526 nan 0.2306 -0.1202
## 440 49.4758 nan 0.2306 -0.1358
## 460 48.1515 nan 0.2306 -0.1350
## 480 47.0966 nan 0.2306 -0.1003
## 500 46.2223 nan 0.2306 -0.2348
## 520 45.2569 nan 0.2306 -0.1366
## 540 44.5323 nan 0.2306 -0.1584
## 560 43.7243 nan 0.2306 -0.1089
## 580 42.9962 nan 0.2306 -0.0628
## 600 42.3117 nan 0.2306 -0.2880
## 620 41.6916 nan 0.2306 -0.0902
## 640 41.0948 nan 0.2306 -0.1138
## 660 40.4783 nan 0.2306 -0.2369
## 680 39.8809 nan 0.2306 -0.0643
## 700 39.4089 nan 0.2306 -0.0395
## 720 38.8299 nan 0.2306 -0.1027
## 740 38.4079 nan 0.2306 -0.2828
## 760 37.8741 nan 0.2306 -0.1307
## 780 37.4330 nan 0.2306 -0.1424
## 800 37.0423 nan 0.2306 -0.1443
## 820 36.5754 nan 0.2306 -0.1626
## 840 36.0294 nan 0.2306 -0.0773
## 860 35.7887 nan 0.2306 -0.1432
## 880 35.4146 nan 0.2306 -0.1236
## 900 35.0992 nan 0.2306 -0.0323
## 920 34.7342 nan 0.2306 -0.0464
## 940 34.4255 nan 0.2306 -0.1259
## 960 34.0454 nan 0.2306 -0.1622
## 980 33.8401 nan 0.2306 -0.0893
## 1000 33.5599 nan 0.2306 -0.0910
## 1020 33.3325 nan 0.2306 -0.0565
## 1040 33.1393 nan 0.2306 -0.1478
## 1060 32.8681 nan 0.2306 -0.1520
## 1080 32.5842 nan 0.2306 -0.1416
## 1100 32.3639 nan 0.2306 -0.0870
## 1120 32.1349 nan 0.2306 -0.1212
## 1140 31.8946 nan 0.2306 -0.1494
## 1160 31.6717 nan 0.2306 -0.1276
## 1180 31.5442 nan 0.2306 -0.1497
## 1200 31.3740 nan 0.2306 -0.1287
## 1220 31.1789 nan 0.2306 -0.0825
## 1240 30.9754 nan 0.2306 -0.1239
## 1260 30.8236 nan 0.2306 -0.0916
## 1280 30.6417 nan 0.2306 -0.1834
## 1300 30.4654 nan 0.2306 -0.1392
## 1320 30.2752 nan 0.2306 -0.0938
## 1340 30.1238 nan 0.2306 -0.1065
## 1360 29.9890 nan 0.2306 -0.0869
## 1380 29.8416 nan 0.2306 -0.0681
## 1400 29.7993 nan 0.2306 -0.1484
## 1420 29.6922 nan 0.2306 -0.0935
## 1440 29.5852 nan 0.2306 -0.1404
## 1460 29.4424 nan 0.2306 -0.1369
## 1480 29.3890 nan 0.2306 -0.1399
## 1500 29.2086 nan 0.2306 -0.0919
## 1520 29.1050 nan 0.2306 -0.1738
## 1540 29.0109 nan 0.2306 -0.0985
## 1560 28.9058 nan 0.2306 -0.1659
## 1580 28.8039 nan 0.2306 -0.1418
## 1600 28.6836 nan 0.2306 -0.1439
## 1620 28.6034 nan 0.2306 -0.1470
## 1640 28.4980 nan 0.2306 -0.1120
## 1660 28.3910 nan 0.2306 -0.1058
## 1680 28.2499 nan 0.2306 -0.1360
## 1700 28.0951 nan 0.2306 -0.1140
## 1720 28.0154 nan 0.2306 -0.0764
## 1740 27.9393 nan 0.2306 -0.1213
## 1760 27.8366 nan 0.2306 -0.1516
## 1780 27.8112 nan 0.2306 -0.0817
## 1800 27.7443 nan 0.2306 -0.1050
## 1820 27.6864 nan 0.2306 -0.1282
## 1840 27.5787 nan 0.2306 -0.0507
## 1860 27.5582 nan 0.2306 -0.0937
## 1880 27.4412 nan 0.2306 -0.1076
## 1900 27.3738 nan 0.2306 -0.0454
## 1920 27.2789 nan 0.2306 -0.0956
## 1940 27.1815 nan 0.2306 -0.1392
## 1960 27.1028 nan 0.2306 -0.0820
## 1980 27.0637 nan 0.2306 -0.1511
## 2000 27.0235 nan 0.2306 -0.1028
## 2020 26.9347 nan 0.2306 -0.1278
## 2040 26.8404 nan 0.2306 -0.0872
## 2060 26.7778 nan 0.2306 -0.1074
## 2080 26.7048 nan 0.2306 -0.1787
## 2100 26.6610 nan 0.2306 -0.1760
## 2120 26.5537 nan 0.2306 -0.1463
## 2140 26.5259 nan 0.2306 -0.2061
## 2160 26.4604 nan 0.2306 -0.1322
## 2180 26.4076 nan 0.2306 -0.1003
## 2200 26.3905 nan 0.2306 -0.1337
## 2220 26.2713 nan 0.2306 -0.1163
## 2240 26.2807 nan 0.2306 -0.1762
## 2260 26.1945 nan 0.2306 -0.1502
## 2280 26.0970 nan 0.2306 -0.0850
## 2300 26.0312 nan 0.2306 -0.0921
## 2320 26.0160 nan 0.2306 -0.1000
## 2340 26.0149 nan 0.2306 -0.1588
## 2360 25.9428 nan 0.2306 -0.1499
## 2380 25.8676 nan 0.2306 -0.0538
## 2400 25.8622 nan 0.2306 -0.1628
## 2420 25.8055 nan 0.2306 -0.2000
## 2440 25.7835 nan 0.2306 -0.1242
## 2460 25.7219 nan 0.2306 -0.0902
## 2480 25.6740 nan 0.2306 -0.1047
## 2500 25.6816 nan 0.2306 -0.2449
## 2520 25.5883 nan 0.2306 -0.1242
## 2540 25.5543 nan 0.2306 -0.2254
## 2560 25.4758 nan 0.2306 -0.1299
## 2580 25.4492 nan 0.2306 -0.1239
## 2600 25.3781 nan 0.2306 -0.0886
## 2620 25.3923 nan 0.2306 -0.1587
## 2640 25.3381 nan 0.2306 -0.1267
## 2660 25.2311 nan 0.2306 -0.0770
## 2680 25.2554 nan 0.2306 -0.0864
## 2700 25.1934 nan 0.2306 -0.1398
## 2720 25.1665 nan 0.2306 -0.0966
## 2740 25.1568 nan 0.2306 -0.1808
## 2760 25.1023 nan 0.2306 -0.0710
## 2780 25.1075 nan 0.2306 -0.1114
## 2800 25.0579 nan 0.2306 -0.1997
## 2820 25.0161 nan 0.2306 -0.0947
## 2840 24.9547 nan 0.2306 -0.1378
## 2860 24.9290 nan 0.2306 -0.1060
## 2880 24.8519 nan 0.2306 -0.1242
## 2900 24.8170 nan 0.2306 -0.1081
## 2920 24.8006 nan 0.2306 -0.1495
## 2940 24.8130 nan 0.2306 -0.1046
## 2960 24.7694 nan 0.2306 -0.1820
## 2980 24.7572 nan 0.2306 -0.1126
## 3000 24.7522 nan 0.2306 -0.1239
## 3020 24.6830 nan 0.2306 -0.1086
## 3040 24.6292 nan 0.2306 -0.1145
## 3060 24.6154 nan 0.2306 -0.1055
## 3080 24.5643 nan 0.2306 -0.0991
## 3100 24.5155 nan 0.2306 -0.1310
## 3120 24.5281 nan 0.2306 -0.1426
## 3140 24.4493 nan 0.2306 -0.1086
## 3160 24.4859 nan 0.2306 -0.0969
## 3180 24.4969 nan 0.2306 -0.2362
## 3200 24.4491 nan 0.2306 -0.0907
## 3220 24.3993 nan 0.2306 -0.1408
## 3240 24.3398 nan 0.2306 -0.0870
## 3260 24.3484 nan 0.2306 -0.1428
## 3280 24.3045 nan 0.2306 -0.1289
## 3300 24.2932 nan 0.2306 -0.0766
## 3320 24.1917 nan 0.2306 -0.1435
## 3340 24.2219 nan 0.2306 -0.1258
## 3360 24.1798 nan 0.2306 -0.0573
## 3380 24.1487 nan 0.2306 -0.1043
## 3400 24.1712 nan 0.2306 -0.0492
## 3420 24.0916 nan 0.2306 -0.0456
## 3440 24.0528 nan 0.2306 -0.0817
## 3460 24.0614 nan 0.2306 -0.0918
## 3480 24.0629 nan 0.2306 -0.2024
## 3500 24.0164 nan 0.2306 -0.1179
## 3520 24.0062 nan 0.2306 -0.1288
## 3540 23.9972 nan 0.2306 -0.1249
## 3560 23.9713 nan 0.2306 -0.0980
## 3580 23.9605 nan 0.2306 -0.1061
## 3600 23.9468 nan 0.2306 -0.1034
## 3620 23.9403 nan 0.2306 -0.0822
## 3640 23.9423 nan 0.2306 -0.0816
## 3660 23.8858 nan 0.2306 -0.1256
## 3680 23.9108 nan 0.2306 -0.0521
## 3700 23.8608 nan 0.2306 -0.0925
## 3720 23.8501 nan 0.2306 -0.1449gb_model$results## shrinkage interaction.depth n.minobsinnode n.trees RMSE Rsquared
## 3 0.5469976 3 9 3489 11.15145 0.8960526
## 1 0.2306400 8 8 3720 10.30007 0.9107268
## 2 0.3895644 3 16 4080 10.50872 0.9071559
## MAE RMSESD RsquaredSD MAESD
## 3 6.882001 0.4931114 0.008629503 0.1895580
## 1 6.056816 0.4177467 0.007095070 0.1247801
## 2 6.438167 0.3027341 0.005237252 0.1097184gb_model$bestTune## n.trees interaction.depth shrinkage n.minobsinnode
## 1 3720 8 0.23064 8gb_model$finalModel## A gradient boosted model with gaussian loss function.
## 3720 iterations were performed.
## There were 81 predictors of which 81 had non-zero influence.Generalization of Error
# Predicting Tc for training/test
pred_test_gb = predict(gb_model, newdata = test_set)
rmse_test_gb <- RMSE(pred_test_gb,test_set$critical_temp)
# Calculates RMSE of training pred
cat("\nGradient Boosting Model: RMSE for the test predictions =", rmse_test_gb)##
## Gradient Boosting Model: RMSE for the test predictions = 10.84253Visualization of Gernerailzed Error
# Visualizing the fit
gb_test <- ggplot() +
geom_point(aes(x = test_set$critical_temp, y = pred_test_gb),
colour = 'deeppink1',alpha=0.5,size=3) +
ggtitle('Gradient Boosting') +
ylab('Prediction') +
xlab('True Value (Tc)') +
theme_minimal() +
geom_abline(colour = "grey80", size = 1)
gb_test
3.4 K-Nearest-Neighbor Regression
KNN is a non-parametric and instance-based learning, and the only tuning parameter in this algorithm is k, the number neighbors (closest training examples) in the feature space. To make a prediction for an unseen datapoint (a query vector), we simply find the k nearest training examples with the smnallest Euclidean distance to it and take the average as output.
knn_model <- train(critical_temp ~ .,
method = "knn",
tuneGrid = expand.grid(k = c(1:5)), trControl = trainControl(method = "cv", number = 10),
data = training_set,
preProc = c("center", "scale","BoxCox"))
knn_model$results## k RMSE Rsquared MAE RMSESD RsquaredSD MAESD
## 1 1 10.82027 0.9028101 5.531605 0.5105083 0.007880347 0.2159699
## 2 2 10.39815 0.9092257 5.457188 0.4352636 0.006771136 0.1920420
## 3 3 10.52456 0.9067717 5.684855 0.3819896 0.006198385 0.2393790
## 4 4 10.77315 0.9021655 5.934506 0.3410619 0.005543605 0.2100993
## 5 5 10.94480 0.8989689 6.091431 0.3629183 0.005911341 0.2088961knn_model$bestTune## k
## 2 2knn_model$finalModel## 2-nearest neighbor regression modelGeneralization of Error
# Predicting Tc for training/test
pred_test_knn = predict(knn_model, newdata = test_set)
rmse_test_knn <- RMSE(pred_test_knn,test_set$critical_temp)
# Calculates RMSE of training pred
cat("\nKNN Model: RMSE for the test predictions =", rmse_test_knn)##
## KNN Model: RMSE for the test predictions = 11.80047Visualization of Gernerailzed Error
# Visualizing the fit
knn_test <- ggplot() +
geom_point(aes(x = test_set$critical_temp, y = pred_test_knn),
colour = 'red',alpha=0.5,size=3) +
ggtitle('K Nearest Neighbor') +
ylab('Prediction') +
xlab('True Value (Tc)') +
theme_minimal() +
geom_abline(colour = "grey80", size = 1)
knn_test
4. Model Comparison
model <- c("Liner_regression",
"Lasso_regression",
"Linear_reg_featureCrosses",
"Random_forest",
"Gradient_boosting",
"KNN"
)
rmse <- c(round(rmse_test1.3,4),
round(rmse_test2.2,4),
"HUGE (Overfitting)",
round(rmse_test_rf,4),
round(rmse_test_gb,4),
round(rmse_test_knn,4))
models.test.metrics <- data.frame(model,rmse)
models.test.metrics## model rmse
## 1 Liner_regression 18.4647
## 2 Lasso_regression 16.5828
## 3 Linear_reg_featureCrosses HUGE (Overfitting)
## 4 Random_forest 10.2137
## 5 Gradient_boosting 10.8425
## 6 KNN 11.8005Putting them all together, we can tell that the non-linear models clearly did a better job in general. At least the relationship between the true values and the predictions is much stronger. Unlike our linear regression models, there isn’t a gap or a shift in between the Y=X trend.
plot_results <- grid.arrange(Linear_regression_test + theme(plot.title = element_text(size = 10),
axis.title.x = element_text(size=7),
axis.title.y = element_text(size=7)),
Lasso_regression_test + theme(plot.title = element_text(size = 10),
axis.title.x = element_text(size=7),
axis.title.y = element_text(size=7)),
featureCrosses_test + theme(plot.title = element_text(size = 10),
axis.title.x = element_text(size=7),
axis.title.y = element_text(size=7)),
rf_test + theme(plot.title = element_text(size = 10),
axis.title.x = element_text(size=7),
axis.title.y = element_text(size=7)),
gb_test + theme(plot.title = element_text(size = 10),
axis.title.x = element_text(size=7),
axis.title.y = element_text(size=7)),
knn_test + theme(plot.title = element_text(size = 10),
axis.title.x = element_text(size=7),
axis.title.y = element_text(size=7)),
nrow=2)
5. Conclusion
In the beginning we explore the relationships between features and the label. We observed correlations between some of the features, however, there seemed to be no obvious linear relationship between features and the label. So then it makes sense that the linear regression models perform poorly while other non-linear models such as Random Forest and Gradient Boosting outperformed its counterpart using clever methods like bagging and boosting. Even thought we tried encoding the nonlinearity and got a better predictive power for the linear regression, we still suffered from overfitting. Finally, we saw that KNN, the non-parametric model, performs pretty well on the test data, just a little worse than those ensenble learning models. To sum up, Random_forest was the best model while the overfitting linear regression (with thousands of feature crosses) was the worst.
5. Discussion
We have gone through different scenarios in this project. We first tried to improve the linear regression models through feature engineering, feature transformation, and then we dealth with an overfitting model that had thousands of engineered features. Finally, we realised non-linear models, even with raw input data, significantly outperformed any kind of linear regression models. It tells us that a non-linear problem requires a non-linear solution.
Also, all three of the non-linear models performed well with just the full set of raw features, we can still expect them to perform maybe slightly better with the pre-processed features. For example, for algorithms that rely on calculating distance for classification or regression such as KNN, if the feature space is too sparse or too large, calculating the Euclidean distance and searching for neighbors becomes inefficient and this algorithm may suffer from the curse of dimensionality where all vectors are almost equidistant to the search query vector. In this case, pre-processing steps such as feature extraction and dimension reduction performed on the full size raw data can help avoid the effects of the curse of dimensionality.
Reference
- https://machinelearningmastery.com/feature-selection-with-the-caret-r-package/
- https://data.library.virginia.edu/diagnostic-plots/#targetText=Scale%2DLocation,equally%20(randomly)%20spread%20points.
- https://www.coursera.org/learn/ml-regression/home/welcome
- https://www.youtube.com/watch?v=ctmNq7FgbvI
- https://en.wikipedia.org/wiki/Gradient_boosting
- https://www.nature.com/subjects/superconductors